Clinical data on 4-aminoquinones and SARS-CoV-2
In the early stages of the pandemic, the National Health Commission of China incorporated 4-aminoquinones into the treatment guidelines of COVID-19 after Chinese preliminary data demonstrated tolerability and efficacy of chloroquine in treating COVID-19.[2] The expectations for the effectiveness of these medications sharply increased after Gautret et al. published results from their small, open-label, non-randomized clinical trial from a single-center in Marseilles, France.[30] This trial concluded that hydroxychloroquine given at a dose of 600 mg a day in combination with azithromycin was effective in viral load reduction and virus elimination of SARS-CoV-2 as measured by nasopharyngeal PCR in patients across disease severity. The article received criticism for its statistical methods, and particularly for excluding 6 out of the 26 patients who received hydroxychloroquine and were lost to follow up or had a poor outcome. Shortly thereafter, the U.S. Food and Drug Administration (FDA) issued an emergency use authorization that allowed the use of hydroxychloroquine in patients hospitalized due to COVID-19 when clinical trials were not available or participation in clinical trials was not feasible.[31] Other countries, including Brazil, France, Italy, Netherlands, and South Korea issued similar recommendations.
Between March and May, 2020, the period of our trial’s design and recruitment, a myriad of hydroxychloroquine clinical studies were performed. The studies were primarily retrospective analysis reporting on the clinical outcomes of its use in hospitalized patients. In many cases, these studies were subject to confounders and selection biases. Data from randomized controlled trials has more recently become available (Table 1).
A retrospective analysis of hospitalized, non-mechanically ventilated patients across all U.S. Veteran’s Health Administration Medical Centers revealed an association between the use of hydroxychloroquine and an increase in in-hospital mortality.[32] The authors did not specify the excess cause of death in this subgroup, but it did not seem to derive from worse respiratory failure, given that rates of mechanical ventilation were similar between subgroups. An observational study in France comparing 84 hydroxychloroquine treated individuals with 89 patients treated with usual care found similar survival without transfer to the ICU at day 21 in both groups.[33] This trial pre-specified treatment groups prior to hospital admission to minimize selection bias.
A multicenter retrospective study in the United States included 1438 patients from 25 hospitals failed to demonstrate a reduction in in-hospital mortality in patients treated with hydroxychloroquine, azithromycin or both. Combination treatment with hydroxychloroquine and azithromycin was associated with more frequent cardiac arrest (HR 2.13, CI 1.12 – 4.05).[34] A single center study from New York City found no association between hydroxychloroquine and the composite end point of intubation or death (HR 1.04, CI 0.82 – 1.32) in 1376 patients who received usual care or hydroxychloroquine alone.[35] The Henry Ford Hospital Study, a retrospective cohort study of 2541 patients, reported a mortality benefit in patients receiving hydroxychloroquine.[36] This study has been criticized for potential confounding factors including the more frequent use of corticosteroids and tocilizumab in hydroxychloroquine treatment arms, with corticosteroids now recognized as an effective form of treatment in hospitalized patients requiring supplemental oxygen.[37]
A retrospective analysis of a multinational registry including over 96,031 subjects from 671 hospitals in six continents comparing four groups, chloroquine or hydroxychloroquine with and without a macrolide, to usual care reported an increase in hospital mortality and new ventricular arrhythmias with use of hydroxychloroquine and chloroquine with or without macrolide.[38] This study was soon retracted due to questions over data integrity as a significant amount of data provided by a private company, Surgisphere, appeared to have large discrepancies and was not released for independent review citing contractual limitations.
In a retrospective study of critically ill, mechanically ventilated patients with COVID-19, hydroxychloroquine was associated with a substantial decrease in mortality (18.8% in the hydroxychloroquine group versus 47.4% in the control group) and interleukin-6 concentrations.[39] The authors did not report criteria for hydroxychloroquine use and selection bias is an important concern when interpreting these results. Preliminary results of the randomized controlled multicenter RECOVERY trial comparing 1561 patients treated with high-dose hydroxychloroquine to 3155 patients on usual care revealed hydroxychloroquine was not associated with reductions in 28-day mortality but rather with increased length of hospital stay and increased risk of progression to invasive mechanical ventilation.[40]
After review of these trials, the U. S. FDA revoked its Emergency Use Authorization of chloroquine and hydroxychloroquine to treat hospitalized COVID-19 patients. Subsequently, the World Health Organization discontinued the SOLIDARITY trial of hydroxychloroquine treatment in hospitalized COVID-19 patients after interim results revealed little or no reduction in mortality and concerning safety signals, albeit without “solid evidence of increase in mortality.”
A randomized controlled trial addressing mild to moderate hospitalized COVID-19 patients also failed to show that hydroxychloroquine alone or in combination with azithromycin improves clinical status compared to usual care.[41] The first randomized multicenter clinical trial evaluating the effect of hydroxychloroquine (n=75) in comparison to usual care alone (n=75) in mild to moderate COVID-19 did not show differences in the probability of SARS-CoV-2 PCR negative conversion or alleviation of symptoms by day 28 of follow up.[42] The median of 16 days delay between onset of symptoms and randomization could have biased the results towards the non-intervention group.
A recent randomized, double-blind, placebo-controlled clinical trial for treatment of symptomatic non-hospitalized patients with COVID-19 investigated a higher dose of hydroxychloroquine (800mg plus 600mg six to 8 hours later and 600mg daily for 4 days). This trial demonstrated that hydroxychloroquine did not substantially reduce symptom severity in outpatients with early, mild COVID-19. Adverse events were significantly higher in hydroxychloroquine group with gastrointestinal symptoms being most commonly reported. Of note, some limitations included only 58% of patients had laboratory confirmation of SARS-CoV-2 due to testing shortages.[43]
Conflicting results from small clinical trials and a retrospective study on the benefits of hydroxychloroquine with and without azithromycin have been reported. [30, 44-48] Such studies were frequently unpowered and lacked a comparison group.[44, 45, 48] Moreover, the inclusion of individuals with mild or no symptoms and characteristics associated with better prognosis support the evidence that the prognosis in mild to moderate COVID-19 is overall good and treatment is unlikely to benefit these patients.[49]
The use of hydroxychloroquine as a prophylactic agent has been less studied. Through a creative recruitment process utilizing internet-based self-referral and online follow up surveys, Boulware et al.randomized asymptomatic adult individuals from Canada and the U. S. who had a high or moderate-risk exposure to confirmed cases of COVID-19 to hydroxychloroquine or placebo.[50] Hydroxychloroquine did not prevent illness compatible or confirmed to be COVID-19 infection and was associated with increased mild adverse reactions. In light of the overall younger (median age 40) and healthier population recruited, mostly women (51.6%), the question remains if more at risk populations would benefit from the intervention.