Hydroxychloroquine and COVID-19
What do the clinical trials say?
As medical professionals scramble to determine effective treatments for the novel coronavirus that emerged late last year, one potential drug has attracted strong praise and criticism, often finding itself at the base of conspiracy theories and political arguments. Too often politicization can lead to strong opinions that are not grounded in evidence. But what does the actual data say? Here, I present my analysis and overall conclusion of several clinical trials and studies that have examined COVID-19 patients treated with hydroxychloroquine to determine its therapeutic potential. I will let you decide if the growing medical consensus against this therapy is warranted.
Background
Hydroxychloroquine (HCQ) has been used for decades to treat malaria and rheumatic disease. In the early stages of the pandemic, there was hope that HCQ could also be used as a treatment for COVID-19, as several preliminary in vitro studies suggested it may have a positive effect. Based on these and other studies, many different clinical trials were initiated to determine the efficacy of HCQ in treating COVID-19. I have reviewed the literature in search of clinical trials that aimed at determining the efficacy of HCQ in the context of COVID-19 that have posted preliminary or complete results. Though it is possible that I have missed some trials, I tried to be as comprehensive as I could with the list of actual trials that I found. A few words about the types of clinical trials are necessary before we look at the results.
The gold standard of clinical trials is a randomized controlled trial. It is even better when the trials are blinded, which means that the patients or the patients/doctors are not told who receives the treatment and who receives the placebo. However, these trials are expensive and more difficult to get approved. Observational trials can also be initiated to study the effects of a specific treatment, but there is less control over variables and more potential for bias. Retrospective cohort studies are less expensive and easier to carry out due to the fact that the data has already been collected and is more readily accessible. However, these studies are also prone to substantial bias and it is very difficult, if not impossible, to control for any confounding variables. Due to these limitations, the results from observational and retrospective studies should be accepted cautiously/tentatively until large, prospective clinical trials can be conducted.
With these things in mind, I have organized my analysis by the type of study, starting with the gold standard and moving on to less definitive studies. Further, I limited my analysis to published results (as opposed to pre-prints that have not been peer-revied) in all cases except for the results of the RECOVERY study, the reason for which is noted below. There are also several preprints not mentioned here because they have not been peer-reviewed, but their results agree with and strengthen the overall consensus.
Prospective Randomized Controlled Clinical Trials
I was able to identify four randomized controlled trials with published results. None of these four trials showed a significant therapeutic benefit for HCQ when treating patients with COVID-19. Three of these studies were published in journals. The fourth was the RECOVERY trial, a large study studying several potential COVID-19 therapeutics, which released its HCQ results in a press release on June 5, with a publication forthcoming.
In this study, 150 patients were randomized into two groups, a standard of care + HCQ group and a standard of care control group. The HCQ group received 1200mg of HCQ for three days followed by 800mg HCQ daily for 2–3 weeks. Treatment was initiated within 24hrs of randomization. The study outcomes looked to see if patients had a negative conversion of the virus by 28 days and if there were any clinical improvements by 28 days.
Results: The study found no significant difference for virus elimination or improved clinical outcomes between the HCQ and the standard of care groups. Adverse events (diarrhea, vomiting, nausea, etc.) were more prevalent in the HCQ group than the standard of care group.
Trial 2: A Randomized Trial of Hydroxychloroquine as Postexposure Prophylaxis for Covid-19
This study was designed to test the effectiveness of HCQ in preventing symptomatic COVID-19. 821 asymptomatic patients who had been exposed to to the virus were identified and randomized to either take HCQ or a placebo as a prophylactic measure to prevent symptomatic disease. The HCQ dosing regimen was 800mg once, then 600mg 6–8hrs later, filled by 600mg a day for five days. The primary outcome analyzed was symptomatic disease onset.
Results: The study found no significant differences found between the HCQ and the control group, indicating that HCQ had no effect in preventing symptomatic disease after exposure. Side effects were also more common in the HCQ group. This trial is interesting because the participants were given HCQ very early on prior to disease presentation in an effort to prevent symptomatic disease altogether. One of the main criticism of HCQ studies that find no effect is that treatment was not started early enough. This criticism would not apply to this trial.
Trial 3: Hydroxychloroquine with or without Azithromycin in Mild-to-Moderate Covid-19
This trial followed three groups in a 1:1:1 ratio: standard of care (control), standard of care plus HCQ, and standard of care plus HCQ and Azithromycin. HCQ was given at a dose of 400mg twice daily, and Azithromycin was given at a dose of 500mg once daily. The primary outcome was a seven-level clinical condition scale assessed at 15 days of hospitalization.
Results: Neither treatment groups, HCQ or HCQ and Azithromycin, showed any significant clinical benefits over the control group. Negative effects for the heart and liver were more frequent in HCQ groups. HCQ did not improve clinical outcomes by 15 days of hospitalization.
Trial 4: RECOVERY — Randomised Evaluation of COVID-19 Therapy
The Randomised Evaluation of COVID-19 Therapy (RECOVERY) study is a large clinical trial initiated by the University of Oxford that aims to study several potential treatments for COVID-19, including Dexamethasone, Azithromycin, Tocilizumab and convalescent plasma (antibodies from the plasma of COVID-19 patients). I include this trial last because the results have not been released in a peer-reviewed paper yet. However, the results for the HCQ arm of this study were analyzed and released in a press briefing. Whereas a press briefing is not the definitive standard for result publication, there is no reason to doubt the results given here. Further, these results will be released in a forthcoming publication.
Results: This trial was very large, with 1,542 patients randomized to receive HCQ and 3,132 patients to receive the standard of care (control). The primary endpoint was mortality. There was no significant difference in patient mortality (25.7% in HCQ group compared to 23.5% in the control group) or in beneficial effects on hospital stay.
Small, non-randomized prospective trials
The four large randomized trials listed above came to the unanimous conclusion that HCQ did not provide any benefit when used in therapy for COVID-19. However, the results of smaller trials are less clear, as might be suspected due to the limitations that come with small sample size. Here I analyize the results of two smaller, non-randomized trials with conflicting results. Both are from French groups, and the second study was conducted after the first study suggested a beneficial effect of HCQ.
This first study was one of the earliest reports on the efficacy of HCQ and one of the major reasons that the scientific community though HCQ might have a potential benefit in treating COVID-19. In this study, 20 patients received HCQ (200mg three times a day for 10 days) and 16 patients were in a control group. A third group (n=6) was formed after the study was started as some patients in the HCQ group also were given azithromycin. The primary outcome was clearing of the virus at day 6. Secondary outcomes were virus clearance and clinical follow up over the course of the study.
Results: The study did find a significant difference for viral load reduction in patients given HCQ and HCQ + azithromycin. However, it should be noted that the methodology of this study has been criticized and reanalyzed by other groups, showing no significant differences between the groups. This group also did another observational study that claimed that HCQ had a beneficial effect for COVID-19 patients, however they did not include a control which makes analysis of the beneficial effect of the drug over the standard of care impossible.
This study (also in France) was initiated after the positive results of the study above were released. This study was prospective and followed the outcomes of 11 consecutive patients who received HCQ (600mg daily for 10 days) and azithromycin (500mg on the first day, then 250mg daily for 5 days). The outcomes of the study was clearing of the virus to mirror the study above.
Results: This study could not replicated the results of the study above, finding no evidence of strong antiviral activity or clinical benefit of HCQ and azithromycin. It should be noted that there was no control group because the design of the study was to test the results of the study above.
Retrospective and Observational Studies
While not the gold standard, retrospective and observational studies can give us some idea of how well a certain treatment might work. However, we should always be careful in our weighting of these studies, as they are more prone to bias and confounding variables. Here I present five studies, four of which found no evidence of HCQ benefits and one which did find some evidence for HCQ benefit, organized by date of acceptance/online publication.
This study observed 181 patients who required oxygen but not intensive care. The primary outcome measured was survival without transfer to the ICU at day 21. Secondary outcomes measured included overall survival, weaning from oxygen and discharge from hospital. The study followed two groups, the HCQ group receiving a daily dose of 600mg HCQ (84 patients received it within 48hrs and 8 after 48hrs, for a total of 92 patients) and the control group receiving no HCQ (total of 89 patients). The main analysis compared only the HCQ group that received treatment within 48hrs of admission.
Results: No significant differences were found between the HCQ group and the control group for any of the measured outcomes (survival rate, survival rate without severe disease, weaning from oxygen, and hospital discharge). 8 patients in the HCQ groups had heart complications that required them to stop HCQ treatment.
This study looked at data from 1,438 hospitalized patients across multiple centers in the New York metropolitan region who received HCQ (n=271), azithromycin (n=211), both (n=735), or none (control, n=211). The median time to HCQ treatment was one day following admission. Dosing varied among patients. The primary outcome was in-hospital mortality.
Results: The study found no significant differences in in-hospital mortality in any of the treatment groups compared to the control group. The study also found no significant differences in abnormal ECG readings (heart issues) for any of the groups.
Study 3: Observational Study of Hydroxychloroquine in Hospitalized Patients with Covid-19
This study followed a group of 1,376 patients in a large medical center in New York City to examine the association of HCQ treatment and intubation or death. The primary endpoint was timing from the study baseline to intubation or death. 811 patients received HCQ and 565 did not. The suggested HCQ dose at the hospital was 600mg twice on day one, followed by 500mg daily dose on days 2–5. Azithromycin was given at 500mg on day one and then 250mg for the following 4 days.
Results: The results of this study found that there was no significant difference in the risk of intubation or death among the study groups. A secondary analysis of patients who received azithromycin also showed no significant difference. There was no beneficial evidence of HCQ treatment.
Study 4: Outcomes of Hydroxychloroquine Usage in United States Veterans Hospitalized with COVID-19
This was a retrospective study that evaluated 807 patients in the US Veterans Health Administration medical centers. There were three groups identified, patients who received HCQ (n=198), HCQ + azithromycin (n=214), and a control group (n=395) that received neither treatment. Dosing regiments ranged across patients.
Results: This study found that the risk of death from any cause was higher in the HCQ group, but not the HCQ + azithromycin group, when compared to the control group. There was no significant difference in the risk of ventilation and subsequent death among groups. HCQ showed no significant reduction in mortality or a reduced need for ventilation in COVID-19 patients, with or without azithromycin.
This retrospective observational study examined 2,541 patients in the Henry Ford Health System in Southeast Michigan and divided them into four groups: HCQ, HCQ + azithromycin, azithromycin, and neither medication. The primary outcome was in-hospital mortality. HCQ was given at a dose of 400mg twice on day one, followed by a 200mg dose twice daily on days 2–5. Azithromycin was given once at 500mg on day one, and subsequently at 250mg once daily for the next 4 days.
Results: This study did find a significant difference in the HCQ (13.5% mortality) and HCQ + azithromycin (20.1% mortality) groups compared to the control (26.4% mortality) for a reduction of COVID-19 associated mortality. A lower mortality rate was also found for the azithromycin alone group (22.4% mortaility) compared to control. It should be noted that the results here have been criticized due to the fact that the HCQ groups were more likely to also receive steroids, a treatment that has shown some effect in COVID-19 patients. This is reflected in Table 1 of the paper.
Conclusion
Here I have presented the results from several relevant trials of HCQ treatment for COVID-19 patients. Each of these studies has advantages and limitations. Though I tried to be comprehensive, there is a chance that I missed a trial or two; however, I think the above trials are sufficient to draw some tentative conclusions, barring any unexpected results from future HCQ trials (which are ongoing). Though some limited early trials showed potential for HCQ treatment, the likelihood that HCQ is a good treatment for COVID-19 decreased markedly as more data came out. Only one trial since the initial small trials has shown any benefit, and even this trial did not show HCQ to be a miracle drug as many conspiracy theories claim. Further, the results from the gold standard randomized clinical trials have so far been unanimous in determining that HCQ has no beneficial effect on COVID-19. The weight of the data is currently against the drug, and we would do well to reallocate our resources towards other drugs and therapies that show better therapeutic potential.
