The treatment of viral pneumonia is supportive care. The arrival of COVID-19 has not changed that basic principle. Due to the scope of the crisis, there is massive interest in finding other possible treatments. I will review the evidence for those various treatments, starting today with chloroquine and hydroxychloroquine, but I think we will end up back where we started: the treatment of viral pneumonia is supportive care.
By now, because of widespread media coverage and the antics of a certain politician, everyone has heard that chloroquine might cure COVID-19. But what does the evidence really tell us?
There is some laboratory (in vitro) evidence that chloroquine could be effective in inhibiting SARS-CoV-2 (the virus that I will from now on refer to by its clinical name, COVID-19). (Wang 2020) However, in vitro data needs to be treated incredibly carefully. Chloroquine has been shown to have in vitro activity against influenza, but there is a double blind RCT showing no effect in patients (for prevention). (Paton 2011) It also has in vitro effects on Dengue, but no clinical benefit. (Tricou 2010) Similar in vitro effects have been shown against ebola, and again were not replicated in any in vivo animal studies. (Touret 2020)
There are at least 16 clinical trials that have been done in China (although only with a total of 100 patients so far). This includes RCTs, such as this one, but as far as I can tell, none of them are actually finished. People are relying on a research letter and a news briefing that reported that in the initial patients enrolled in these trials, chloroquine was superior to control in “inhibiting the exacerbation of pneumonia, improving lung imaging findings, promoting a virus negative conversion, and shortening the disease course.” (Gao 2020) This sounds promising, but it is hard to make any treatment recommendations based on a press briefing. None of the scientific details are available to review. We should wait to see the published results.
As far as I can tell, there is only one clinical trial published at this point:
PLEASE NOTE: This trial has since “received a statement of concern”, as the journal says it “does not meet the [International Society of Antimicrobial Chemotherapy’s] expected standard”
Gautret P, Lagier J, Parola P, et al. Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial. International Journal of Antimicrobial Agents. 2020; [article]
This is an observational study of patients 12 years and older with documented COVID-19 admitted to a single hospital in France, with a control group from other hospitals around France. They excluded patients with an allergy or contraindication to chloroquine or hydroxychloroquine, G6PD deficiency, QT prolongation, pregnancy, and breastfeeding. The patients were pretty healthy, with only 20% having pneumonia, and 17% being asymptomatic. At the primary centre, all patients were given 200 mg of oral hydroxychloroquine three times a day. Patients at the other centers did not receive hydroxychloroquine. (This is clearly not an ideal control, as there are lots of reasons that care might differ between hospitals, especially when one of those hospitals is organizing a trial, while the others aren’t.) They originally enrolled 42 patients. However, 6 were excluded from the main site, mostly because of death or transfer to the ICU, while no patients were excluded from the controls. (They didn’t do an intent-to-treat analysis). This means the sickest patients were excluded from the treatment group, so we should expect their results to look better. The patients were not similar at baseline (most of the controls didn’t have the quantitative PCR test done at enrollment and the demographics of the groups were significantly different), and they also weren’t treated the same throughout the trial (eg. the treatment group was swabbed more often than the no treatment group). In other words, there are many sources of bias in this trial.Their primary endpoint was viral clearance on day 6, which occured in 70% of the treatment group and 13% of the controls (p=0.001). Some patients also happened to get azithromycin, and those patients happened to have 100% viral clearance on day 6. They make no mention of harm or adverse events in the paper (although there are already multiple reports of people overdosing on these dangerous medications because of widespread reporting that it might help with COVID-19. See here, here, and here.)
This study is a mess, and clearly should not drive clinical practice. It is the kind of study we wouldn’t even notice in normal times, but the pandemic is causing people to throw out scientific principles. This is an observational trial with significant sources of bias. It doesn’t assess any patient oriented or clinical outcomes. And it doesn’t look at the group of patients we are most concerned about (they included asymptomatic patients here). At best, these results represent and an association at high risk of bias. The study raises a hypothesis. When combined with other data, it might be reasonable to proceed with an RCT, but based on historical examples, changing clinical practice at this point is more likely to do harm than good. With regards to azithromycin, the results are so weak I am not sure it’s even worth considering it a hypothesis.
There is no reason to be using chloroquine at this point. I will update the blog when the Chinese RCTs are published. There is also a very large RCT underway through the WHO (SOLIDARITY) and many other RCTS (eg NCT04303507 and NCT04308668).
There is a systematic review published, but it only includes “one narrative letter, one in-vitro study, one editorial, expert consensus paper, two national guideline documents”. (Cortegiani 2020) They conclude that there is reasonable pre-clinical justification for a clinical trial, but that chloroquine and hydroxychloroquine should be considered experimental, and there is no justification for widespread clinical use.
A pandemic is not an excuse for sloppy science. It is ridiculous that pharmacies are out of these drugs, so that they are not available to patients who actually need them. It is ridiculous that patients have died because they thought this drug might prevent COVID-19, fueled by irresponsible media coverage and unscientific statements from a world leader. Science still applies in 2020. We need to do better.
March 29, 2020 Update:
Gautret P, Lagier J, Parola P, et al. Clinical and microbiological effect of a combination of hydroxychloroquine and azithromycin in 80 COVID-19 patients with at least a six-day follow up: an observational study. Prepublished online
This is a follow up study from the same group as our first paper. I don’t believe it is published in a journal yet, but it is available online here. This is even lower quality evidence than the last paper, because they don’t include a control group. They just tell us about 80 COVID-19 patients in whom they gave a combination of hydroxychloroquine and azithromycin. They did include some of the same patients as the first paper. The patients are young (median age 52), and sound relatively health, with only half having lower respiratory tract involvement, and only 15% being febrile, and 94% having a low NEWS score. Despite that, only 81% have a “favourable outcome” at the time the paper was written up. One patient was admitted to the ICU and 1 patient died. Without a control group, it is impossible to know what this data means, but overall those numbers don’t sound good, and could easily be consistent with this treatment combination harming patients rather than helping them.
March 30, 2020 Update:
A pilot RCT was published. It is not in English, it is not in PubMed, and the listed DOI doesn’t match when I look it up, so I don’t know if it is officially published yet, or just pre-released online. The paper can be found here. They randomized 30 COVID-19 patients to either get hydroxychloroquine or standard care (unblinded) for 5 days. It didn’t work. More people had negative viral swabs in the control group (93%) than the treatment group (87%), and the only patient who deteriorated and needed ICU was in the treatment group. Thus, based only on the abstract, the first RCT is negative, although the numbers are tiny, and I can’t fully appraise the paper because I am monolingual.
April 2, 2020 Update
Just a small note to remind everyone of the significant publication bias that can come with observational data. People choose when they want to publish observational data, and in general, they are not going to publish unexciting results. Look at this tweet from Dr. Roberto Cosntini in Italy. They routinely gave hydroxychloroquine to all patients admitted with COVID-19, and their results were far from miraculous. (In fact, people are still struggling to figure out why the mortality rate in Italy was the highest in the world). This data will probably never be published, but this is thousands of patients, as compared to the less than 100 patients everyone is talking about in France.
April 3, 2020 Update
These is another RCT available in pre-publication format (available here). I am concerned about these prepublication documents, because they do seem to change with time. (The one I linked to is “v2”, while the one I originally appraised was “v1”.) This is a non-blinded, single-center RCT that included 62 adult patients positive for COVID-19, with CT findings of pneumonia, and low oxygen requirements (mild disease). Their endpoint is “time to clinical recovery”, meaning afebrile and no cough for 72 hours, which is clearly not the outcome we care most about in this pandemic, and is somewhat subjective (remember, there is no placebo). Fever was shorted by 1 day with treatment (3.2 days vs 2.2 days, p=0.0008). Cough was shorted by 1 day with treatment (3.1 days vs 2.0 days, p=0.0016). There were no deaths. Aside from the the fact that this is a small, single center, open-label trial, with a subjective outcome of questionable importance, there are some major red flags here. The registration for this trial suggests there were supposed to be 3 groups, they were supposed to enroll 300 patients, and there was supposed to be a placebo. What happened to the patients taking the lower dose hydroxychloroquine? What happened to the placebo? Why did they stop the study early? They state that their randomization was “stratified by site”, but it was a single center study, so that doesn’t make sense. There are way too many problems with this trial to put a lot of faith in the outcomes. Either way, as we are filling up our ICUs, 1 day less of cough is not the outcome I am looking for. This RCT, although the first ever “positive trial”, is in no way a game changer for me.
April 30, 2020 Update
There have been a few new studies since I last had the chance to update this point.
Tang (2020 – available as preprint here) in a multi-center, open-label trial of 150 hospitalized adult patients with mild COVID in China. They gave hydroxychloroquine 1200 mg for 3 days and then 800 mg for 2 weeks and compared that to standard care. There was no difference in the primary outcome of negative seroconversion from SARS-CoV-2 within 28 days. There was also no difference in symptoms. There was, however, a significant increase in adverse events (30% vs 9%, p=0.001). In other words, hydroxychloroquine is harmful.
There is also a chart review looking at 368 patients admitted to a VA hospital in the United States. (Magagnoli 2020) The rate of death was higher in patients who were treated with hydroxychloroquine than in those who were not (28% vs 11%, p=0.03). Of course, this is not a randomized trial, and hydroxychlorquine might just have been given to sicker patients, but it is not promising data.
Borba (2020) is a double blind RCT out of Brazil comparing two doses of chloroquine (600 mg BID for 10 days or 450 mg BID on day one then daily for 4 more days). They were supposed to enrol 440 patients, but the trial was stopped early after only 81 because mortality was significantly higher in the high dose chloroquine group.
May 4, 2020 Update
There is a systematic review in Academic Emergency Medicine. (Chowdhury 2020) Their conclusion is “there is currently not enough data available to support the routine use of HCQ and CQ as therapies for COVID-19.” That won’t be surprising to anyone based on the above, but I include to mention that they found 29 ongoing registered trials looking at this issue. Although we want an answer to this question, 29 trials seems a little excessive when there are so many other important, unanswered questions in medicine.
May 7, 2020 Update
Another large observational study has shown no benefit from hydroxychloroquine. Of 1376 patients admitted to a single hospital in New York City, 811 (59%) receive hydroxychlorquine. The hydroxychloroquine was used in sicker patients, and they therefore did worse overall, but when they do propensity matching, there was no difference in the primary outcome of time to death or intubation (hazard ratio, 1.04; 95% CI, 0.82 to 1.32). (Geleris 2020)
May 22, 2020 Update
It is hard to keep with all the data that is published (while still working and writing about things that actually matter). It is pretty clear that this point that hydroxychloroquine isn’t going to provide a huge benefit. What is becoming more and more clear is that it might actually be harmful. In a new registry study that included 96,032 COVID patients, of whom 1868 received chloroquine, 3783 received chloroquine with a macrolide, 3016 received hydroxychloroquine, and 6221 received hydroxychloroquine with a macrolide, and 81,144 received none of the combination, all the treatment combinations were independently associated with increased mortality and an increased ventricular arrhythmias, after controlling for multiple confounders. We are again faced with the limitations of observational data, but there isn’t even a hint of benefit here. Clearly, aside from the RCTs that are underway, no one should be using hydroxycholoquine or chloroquine to treat or try to prevent against COVID. (Mehra 2020)
This paper has been retracted among rumors of fraud.
June 3, 2020 Update
There is a new RCT looking at hydroxychlorquine for prophylaxis, and the quick summary is that it didn’t work. (Boulware 2020) This is a double-blind, placebo-controlled RCT that enrolled 821 asymptomatic people who had a household contact (34%) or an occupational exposure without appropriate PPE (66%) to someone with confirmed COVID-19. They were randomized to either hydroxychloroquine (800 mg once, followed by 600 mg in 6 to 8 hours, then 600 mg daily for 4 additional days)(800 mg once, followed by 600 mg in 6 to 8 hours, then 600 mg daily for 4 additional days) or placebo. In an interesting twist, enrollment was primarily through social media, and the drugs were shipped by overnight delivery from one of two pharmacies. The primary outcome was symptomatic illness, and occurred in 11.8% of those taking hydroxychloroquine and 14.3% of those taking placebo (ARR -2.4%; 95% CI -7 to 2.2%; p=0.35). Of course, the primary outcome is 1 step removed from the most important outcome, which is a bad outcome from COVID, but limiting the spread of disease would also be valuable. There was only 1 hospitalization in both groups and no deaths. Most patients did not have laboratory confirmed COVID, but rather just symptomatic disease consistent with COVID. Only 75% of the hydroxychloroquine group completed their full course, as compared to 83% of the placebo group, which could clearly limit efficacy. Side effects occured in 40% of the hydroxychloroquine as compared to 17% of the placebo group.
Bottom line: This RCT demonstrated no benefit and increased harm from using hydroxychloroquine for COVID prophylaxis. Although we cannot completely exclude a small benefit, there is very little reason to believe one would exist based on what we know so far.
July 2020 Update
It is barely news at this point. The preprint for the hydroxychloroquine part of the RECOVERY trial is now published, with 4716 patients randomized to either hydroxychloroquine or usual care. There was no difference – mortality was 25% in both groups. By now it is pretty clear, this drug has no role in the mangement of COVID.
There is another multi-centre open-label RCT (Cavalcanti 2020) that randomized 667 patients, and neither the hydroxychloroquine or hydroxychloroquine plus azithromycin groups were better than standard care.
Borba MGS, Val FFA, Sampaio VS, et al. Chloroquine diphosphate in two different dosages as adjunctive therapy of hospitalized patients with severe respiratory syndrome in the context of coronavirus (SARS-CoV-2) infection: Preliminary safety results of a randomized, double-blinded, phase IIb clinical trial (CloroCovid-19 Study). 2020. Not published. Available preprint here.
Boulware DR, Pullen MF, Bangdiwala AS, et al. A Randomized Trial of Hydroxychloroquine as Postexposure Prophylaxis for Covid-19 [published online ahead of print, 2020 Jun 3]. N Engl J Med. 2020;10.1056/NEJMoa2016638. doi:10.1056/NEJMoa2016638 PMID: 32492293 [free full text]
Cavalcanti AB, Zampieri FG, Rosa RG, et al. Hydroxychloroquine with or without Azithromycin in Mild-to-Moderate Covid-19 N Engl J Med. 2020;
Chowdhury MS, Rathod J, Gernsheimer J. A Rapid Systematic Review of Clinical Trials Utilizing Chloroquine and Hydroxychloroquine as a Treatment for COVID-19 [published online ahead of print, 2020 May 2]. Acad Emerg Med. 2020;10.1111/acem.14005. doi:10.1111/acem.14005 PMID: 32359203 [article]
Cortegiani A, Ingoglia G, Ippolito M, Giarratano A, Einav S. A systematic review on the efficacy and safety of chloroquine for the treatment of COVID-19 Journal of Critical Care. 2020;
Gao J, Tian Z, Yang X. Breakthrough: Chloroquine phosphate has shown apparent efficacy in treatment of COVID-19 associated pneumonia in clinical studies BST. 2020; 14(1):72-73. PMID: 32074550
Gautret P, Lagier J, Parola P, et al. Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial. International Journal of Antimicrobial Agents. 2020
Geleris J, Sun Y, Platt J, et al. Observational Study of Hydroxychloroquine in Hospitalized Patients with Covid-19 N Engl J Med. 2020; [article]
Magagnoli J, Narendran S, Pereira F, et al. Outcomes of hydroxychloroquine usage in United States veterans hospitalized with Covid-19. 2020. Not published. Available preprint here.
Mehra RM, Desai SS, et al. Hydroxychloroquine or chloroquine with or without a macrolide for treatment of COVID-19: a multinational registry analysis. Lancet. 2020. DOI: 10.1016/S0140-6736(20)31180-6
Paton NI, Lee L, Xu Y, et al. Chloroquine for influenza prevention: a randomised, double-blind, placebo controlled trial The Lancet Infectious Diseases. 2011; 11(9):677-683.
Tang W, Cao Z, Han M, et al. Hydroxychloroquine in patients with COVID-19: an open-label, randomized, controlled trial. 2020. Not published. Available preprint here.
Touret F, de Lamballerie X. Of chloroquine and COVID-19 Antiviral Research. 2020; 177:104762-.
Tricou V, Minh NN, Van TP, et al. A Randomized Controlled Trial of Chloroquine for the Treatment of Dengue in Vietnamese Adults PLoS Negl Trop Dis. 2010; 4(8):e785-.
Wang M, Cao R, Zhang L, et al. Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro Cell Res. 2020; 30(3):269-271.
Morgenstern, J. Hydroxychloroquine for COVID: No good evidence yet, First10EM, March 26, 2020. Available at: