Does hydroxychloroquine work? Here’s what the studies say so far…
There has been lots of excitement about hydroxychloroquine as a treatment option for COVID-19; early on, this excitement was based on a few small studies and anecdotal reports from physicians. Since then, more studies have come out looking at the effectiveness of hydroxychloroquine in COVID-19 patients. Below is a summary of the study results so far.
For each study, I provide a simple Yes/No answer to did patients who got hydroxychloroquine do better than the patients who didn’t receive it? But this is very much an over-simplification: the type of study, details of the study population, details of the statistical analyses performed, and other strengths and weaknesses of the study should always be taken into account when interpreting the “Yes/No” conclusions (discussion of these is beyond the scope of this post.) Additionally, I only included studies that were (1) done in humans (not in cell lines) and (2) details of the results were made available.
First – a quick refresher on how to evaluate studies. More details are provided at the end of this post, but as you go through these studies, please remember two very important things:
1. Randomized >>> Observational
2. Big Sample Size >>> Small Sample Size
Hydroxychloroquine Study Results
(ordered by sample size) – last updated July 2020
Note: I have done my best to find all the studies to date, but I could have missed some. If you see one I missed, please send it to me and I’ll update this post. Links to studies are provided in every title!
This is by far the best study we have to date, as it is the largest and it is randomized: Effect of Hydroxychloroquine in Hospitalized Patients with COVID-19: Preliminary results from a multi-centre, randomized, controlled trial
Type of Study: Randomized
Outcome: Death within 28 days
Sample Size: 4716 patients
Important notes: not yet peer-reviewed
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: Death within 28 days
Sample Size: 4642 patients
Important notes: not yet peer-reviewed, some patients got azithromycin
Did patients who got hydroxychloroquine do better? No*
*they did see a slight reduction in % of patients discharged with hydroxychloroquine, but no difference in mortality. They also saw a trending increase risk in mortality in hydroxychloroquine + azithromycin group.
Type of Study: Observational
Outcome: Death / Hospital Stay ≥ 10 days / Transfer to ICU / Viral Shedding
Sample Size: 3737 patients
Important notes: patients also got azithromycin, patients who didn’t get drugs were older and sicker
Did patients who got hydroxychloroquine do better? Yes
Type of Study: Observational
Outcome: Death / Need for Mechanical Ventilation
Sample Size: 3,372 patients (though main analysis is 1820 patients)
Important Notes: not yet peer-reviewed, some patients also got azithromycin
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: In hospital mortality
Sample Size: 2541 patients
Important Notes: Some patients also got azithromycin, discussion of confounders here.
Did patients who got hydroxychloroquine do better? Yes
Hydroxychloroquine and Tocilizumab Therapy in COVID-19 Patients – An Observational Study
Type of Study: Observational
Outcome: Death
Sample Size: 2512 patients
Important Notes: Some patients got azithromycin and tocilizumab, not yet peer-reviewed.
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: Death
Sample Size: 1438 patients
Important Notes: Some patients also got azithromycin
Did patients who got hydroxychloroquine do better? No
Observational Study of Hydroxychloroquine in Hospitalized Patients with Covid-19
Type of Study: Observational
Outcome: Intubation or Death
Sample Size: 1376 patients
Important Notes: Patients who got hydroxychloroquine were already sicker prior to treatment.
Did patients who got hydroxychloroquine do better? No
Outcome: Death, Clinical Worsening, Viral Shedding
Sample Size: 1061 patients
Important Notes: No Control Group
Did patients who got hydroxychloroquine do better? No evidence provided (no control group)
Type of Study: Observational
Outcome: Hospitalization
Sample Size: 636 patients
Important Notes: Not all patients were confirmed to have COVID, and stats are weird.
Did patients who got hydroxychloroquine do better? Inconclusive*
*patients were not tested for COVID. also stats are funky in this paper. I do not trust it.
Type of Study: Observational
Outcome: Death
Sample Size: 568 patients
Important Notes: not peer-reviewed
Did patients who got hydroxychloroquine do better? Yes
Low dose of hydroxychloroquine reduces fatality of critically ill patients with COVID-19
Type of Study: Observational
Outcome: Death, measures of inflammation
Sample Size: 550 patients
Important Notes: issues about statistics reported have been raised on pubpeer
Did patients who got hydroxychloroquine do better? Yes
Type of Study: Observational
Outcome: Hospitalization / Death
Sample Size: 518
Important Notes: No way to know if control group was comparable; no details provided
Did patients who got hydroxychloroquine do better? Inconclusive
Outcomes of hydroxychloroquine usage in United States veterans hospitalized with Covid-19
Type of Study: Observational
Outcome: Use of Ventilator and/or Death
Sample Size: 368 patients
Important notes: Not yet peer-reviewed
Did patients who got hydroxychloroquine do better? No
Hydroxychloroquine for Early Treatment of Adults with Mild Covid-19: A Randomized-Controlled Trial
Type of Study: Randomized
Outcome: viral RNA load in nose up to 7 days after treatment start, disease progression, time to complete resolution of symptoms
Sample Size: 293
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: Transfer to ICU and/or Death
Sample Size: 181 patients
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: Death
Sample Size: 166 patients
Important notes: Not yet peer-reviewed
Did patients who got hydroxychloroquine do better? Yes
Type of Study: Randomized Trial
Outcome: Positive for virus after 28 days
Sample Size: 150 patients
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: time to death, ICU admission, or withdrawal of supportive care
Sample Size: 89 patients
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: Positive for Virus, Length of Hospital Stay, Clinical Outcome
Sample Size: 80 patients
Important notes: No control group
Did patients who got hydroxychloroquine do better? No evidence provided (no control group)
Type of Study: Quasi-randomized (I would put this more as observational)
Outcomes: mortality, respiratory status
Sample Size: 63 patients
Important notes: this study was written in French and this is based off the google-translated abstract
Did patients who got hydroxychloroquine do better? Mixed Results (mortality was higher but improvement in respiratory status was better in the hydroxychloroquine group )
Efficacy of hydroxychloroquine in patients with COVID-19: results of a randomized clinical trial
Type of Study: Randomized Trial
Outcome: Time to Clinical Recovery
Sample Size: 62 patients
Important notes: Not yet peer-reviewed
Did patients who got hydroxychloroquine do better? Yes (modest effect)
Type of Study: Randomized Trial
Outcome: Time to Clinical Recovery
Sample Size: 48 patients
Important notes: Not yet peer-reviewed, also included chloroquine group
Did patients who got hydroxychloroquine do better? Yes (modest effect)
Type of Study: Observational
Outcome: Positive for Virus after 6 days
Sample Size: 36 patients
Important notes: major issues with study design
Did patients who got hydroxychloroquine do better? Mehhhhhh*
*the authors conclude there is an effect, but the study design had so many issues that I wrote a whole other post about it.
Type of Study: Observational
Outcome: Positive for Virus
Sample Size: 36 patients
Important notes: I think they may have artificially inflated their statistical power in the way they did their analysis, but I’d have to dig into it more to be sure.
Did patients who got hydroxychloroquine do better? Yes
Type of Study: Observational
Outcome: Positive for Virus
Sample Size: 34 patients
Important notes: not yet peer-reviewed
Did patients who got hydroxychloroquine do better? No (those who got the drug did worse)
Type of Study: Randomized Trial
Outcome: Positive for virus after 7 days of treatment
Sample Size: 30 patients
Did patients who got hydroxychloroquine do better? No
Type of Study: Observational
Outcome: Death (in the hospital)
Sample Size: 96,032 patients
Important Notes: Also includes analysis of chloroquine and macrolides (such as azithromycin)
Did patients who got hydroxychloroquine do better? No (those who got the drug did worse)
STUDY RETRACTED – leaving here for record purposes only.
So, does it work?
Most studies (13 studies) showed no improvement with hydroxychloroquine treatment, including the best designed study we have as well as 6 big studies (>1000 people). However, some studies (8) did show an effect of hydroxychloroquine (though these were often smaller studies, with only 2 studies with >1000 people). How do we interpret this? One way is to go back and do a meta-analysis — where someone pulls all the available data from the studies and re-analyzes the data from multiple studies together. This study did just that, here is what they found:
The Role of Hydroxychloroquine in the Age of COVID-19: A Periodic Systematic Review and Meta-Analysis Sample Size: 21 studies, 103,486 patients
Important Notes: not yet peer reviewed, some patients also received azithromycin
Did patients who got hydroxychloroquine do better? No
Studies of Hydroxychloroquine as Prophylaxis
(ordered by sample size)
prophylaxis: a treatment that is given to prevent a disease from developing (rather than treat a disease that somebody already has)
A Cluster-Randomized Trial of Hydroxychloroquine as Prevention of Covid-19 Transmission and Disease
Type of Study: Randomized Trial
Outcome: PCR-confirmed symptomatic Covid-19 within 14 days
Sample Size: 2314 patients
Important notes: not yet peer-reviewed
Were patients who got hydroxychloroquine less likely to develop COVID? No
A Randomized Trial of Hydroxychloroquine as Postexposure Prophylaxis for Covid-19
Type of Study: Randomized Trial
Outcome: laboratory-confirmed COVID-19, or illness compatible with COVID-19, within 14 days
Sample Size: 821 patients
Were patients who got hydroxychloroquine less likely to develop COVID? No
Type of Study: Observational
Outcome: positive for COVID-19
Sample Size: 106
Important notes: not yet peer-reviewed
Were patients who got hydroxychloroquine less likely to develop COVID? Yes
What do all these words describing studies mean?
Type of Study: All of the studies reported here are some form of either an observational or randomized trial. What’s the difference? An observational study is a study that is done by looking back at what happened organically in the hospital. These studies usually use hospital records to see who got the treatment, who didn’t, and how these groups of patients did over the course of their hospital stay. These studies are important and provide much better evidence than anecdotal reports by physicians, but there are down sides — the main one is that there is no guarantee that those who got the treatment and those who didn’t are truly comparable groups of patients. For example, maybe patients who got hydroxychloroquine were already sicker than those who didn’t — this would effect the differences in outcomes between the treatment and control groups. There are some fancy statistical methods to help correct for these differences, but they are not perfect. Observational studies also are prone to the placebo effect, where patients improve because of the hope of treatment (and possibly other more complicated reasons), and not due to the actual effect of the drug. A randomized study takes care of one of these problems — patients are randomized to either the treatment or control group at the beginning, which usually takes care of differences in populations between the two groups. The control group in these cases is usually the standard-of-care (everything the hospital would normally do for a COVID-19 patient). If it is a randomized placebo-controlled study, then that study also takes care of the placebo effect problem by giving the control group a placebo in addition to the standard-of-care. So if randomized placebo-controlled studies are clearly the best, why don’t we do that for every study? Because they are much harder to do (they take time to get up and running, they cost more, they require much more time on the part of the researchers, etc.)
Outcome: This is simply the patient outcome (positive for virus after 7 days, fever, transfer to ICU, death, etc.) the study used to measure if there was an effect of the treatment. All of these in some way measure how sick the patient is.
Sample Size: How many patients were included in the study (more is always better).
Disclaimer: This post is not intended to provide medical advice or guide treatment decisions. Please consult your physician for questions about medical treatments. COVID-19 treatment guidelines provided by the CDC can be found here.
