Hydroxychloroquine in patients with COVID-19: an open-label, randomized, controlled trial

INTRODUCTION

The coronavirus disease 2019 (COVID-19) caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has swept into more than 200 countries, areas or territories within the four months. As of 5 April, more than 1 million infections and 60 thousands of deaths have been reported.¹

Several agents or drugs including, remdesivir, favipiravir, ribavirin, lopinavir–ritonavir (used in combination) and chloroquine (CQ) or hydroxychloroquine (HCQ), have been highlighted based on the promising in-vitro results and therapeutic experiences from another two coronavirus diseases including the severe acute respiratory syndrome (SARS) and the Middle East respiratory syndrome (MERS)². However, none of these promising results has yet been translated into clinical benefits of patients with COVID-19, including lopinavir–ritonavir, reported from the most recently failed trial.³

Another “wonder drug”, CQ and its hydroxy-analogue HCQ, are glaring on the list of COVID-19 therapy, due to potent antiviral activity against SARS-CoV-2 from in-vitro studies,^4,5 and promising results from news reports of some ongoing trials.⁶ Despite their unclear benefits, CQ and HCQ are both recommended for off-label use in the treatment of COVID-19 by the Chinese National guideline⁷ and recently authorized by the U.S. Food and Drug Administration for emergency uses.⁸ HCQ was also recently recommended by the American president Donald Trump. Such a presidential endorsement stimulates an avalanche of demand for HCQ, which buried the dark-side of this drug. Deaths have been reported in Nigeria among people self-treating for apparent COVID-19 with CQ overdoses.⁹ Retinopathy, gastrointestinal and cardiac side effects are well documented with the use of CQ or HCQ in the treatment of malarial and rheumatic diseases.¹⁰ HCQ is preferred in clinical applications due to its lower toxicity, particularly retinal toxicity,¹⁰ and three times the potency against SARS-CoV-2 infection comparing to CQ in the recent in-vitro study.⁵ Currently, there is no convincing evidence from well-designed clinical trials to support the use of CQ/HCQ with good efficacy and safety for the treatment of COVID-19. Rapidly conduction of such trails with high-quality is challenging in the face of a dangerous coronavirus outbreak, in which, healthcare workers are under overwhelming work and highest risk of exposure to developing COVID-19.¹¹

Having encountered numerous challenges, we conducted a multicenter, open-label, randomized, controlled trial to assess the efficacy and safety of HCQ sulfate in adult patients with COVID-19. A clearer verdict will come from such a trial for the use of HCQ in patients with COVID-19.

METHODS

RESULTS

Patient Baseline Characteristics

A total of 191 patients admitted with COVID-19 from February 11, 2020 to February, 29 2020, were assessed for eligibility, of which 41 did not meet eligibility criteria. The remaining 150 patients underwent randomization; Among them, 75 patients were assigned to SOC and 75 patients to SOC plus HCQ group (Figure 1). The mean age of the patients was 46 years and 55% were male. The mean day from disease onset to randomization was 16.6 and 89% of the patients had concomitant medication before randomization. The majority of the patients had mild to moderate COVID-19 (99%) and only 2 patients (1%) were severe upon screening. Baseline demographic, epidemiological and clinical characteristics of the patients between the two groups are shown in Table 1.

View this table:

• View inline
• View popup

Table 1. Baseline Demographic and Clinical Characteristics of the Patients in the Intention-to-Treat Population.

• Download figure
• Open in new tab

Figure 1.

Screening, Randomization, and Follow-up. Shown is the disposition of the trial participants.

By 14 March 2020 (the cutoff date for data analysis) The median duration of follow-up was 21 days (range, 2 to 33) in the SOC group and 20 days (range, 3 to 31) in the SOC plus HCQ group. Of the 75 patients assigned to receive SOC plus HCQ, 6 patients did not receive any dose of HCQ; of them, 3 patients withdrew consent and 3 patients refuse to be administrated HCQ.

Primary Outcome

Overall, the negative conversion rate of SARS-CoV-2 among patients who were assigned to receive SOC plus HCQ was 85.4% (95% confidence interval [CI], 73.8%, 93.8%), similar to that of the SOC group 81.3% (95%CI, 71.2% to 89.6%) within 28-day. Negative conversion rate at specific time-point, 4-, 7-, 10-, 14- or 21-day was also similar between the two groups. The negative conversion time did not differ between SOC plus HCQ and SOC group (median, 8 days vs. 7 days; hazard ratio, 0.846; 95%CI, 0.580 to 1.234; P=0.341) (Figure 2_panel A).

• Download figure
• Open in new tab

Figure 2.

Kaplan–Meier curves of the time to negative conversion of SARS-CoV-2 on RT-PCR test in SOC plus HCQ versus SOC in the intention-to-treat population and post-hoc subgroup analysis. Shown in panel A are data for 75 patients assigned to SOC plus HCQ and 75 assigned to SOC. The median time to negative conversion in the SOC plus HCQ group (8 days; 95%CI, 5 to 10) was similar to that in the SOC group (7 days; 95%CI, 5 to 8) (P=0.34). Data from patients who did not have negative conversion were censored (tick marks) at the last visit date. Shown in panel B are forest plots of subgroup analyses in which no difference was observed in subgroups regarding virus negative conversion between SOC plus HCQ and SOC treatment. SOC=standard-of-care; HCQ=hydroxychloroquine; CI=confidence interval.

Post hoc analyses were performed in subgroups to explore any decrease of negative conversion time by the addition of HCQ upon SOC. No such effects were observed in the analyzed subgroups according to age (≥45 years versus <45 years), BMI value (≥24 kg/m² versus <24 kg/m²), presence or absence of existing conditions, days between disease onset and randomization (≥7 days versus <7 days), baseline CRP value (≥upper limit of normal versus <upper limit of normal), baseline lymphocyte count (<lower limit of normal versus ≥lower limit of normal) and with or without contaminant use of potential anti-viral agents for treating COVID-19 during the study period (Figure 2_panel B).

Secondary Outcome

The overall rate of symptoms alleviation within 28-day was not different between patients with SOC with (59.9%, 95%CI, 45.0% to 75.3%) and without HCQ (66.6%, 95%CI, 39.5% to 90.9%). The median time to alleviation of clinical symptoms was similar in the SOC plus HCQ group than that in the SOC group (19 days versus 21 days). More rapid alleviation of clinical symptoms with SOC plus HCQ than with SOC alone was observed during the second week since randomization (Figure 3_panel A). The efficacy of HCQ on the alleviation of symptoms (Hazard ratio, 8.83, 95%CI, 1.09 to 71.3) was more evident when the confounding effects of other anti-viral agents were removed in the post-hoc subgroup analysis (Figure 3_panel B). No significant difference between SOC plus HCQ group and SOC group on symptoms improvement was observed in other subgroup analyses (Figure 3_panel B).

• Download figure
• Open in new tab

Figure 3.

Kaplan–Meier curves of the time to alleviation of clinical symptoms in SOC plus HCQ versus SOC in the intention-to-treat population and post-hoc subgroup analysis. Shown in panel A are data for 55 symptomatic patients assigned to SOC plus HCQ and 64 symptomatic patients assigned to SOC. The median time to alleviation of clinical symptoms in the SOC plus HCQ group (19 days; 95%CI, 14 to 22) was similar to that in the SOC group (21 days; 95%CI, 14 to not estimable) (P=0.97). Data from patients who did not have symptoms alleviation were censored (tick marks) at the last visit date. Shown in panel B are forest plots of subgroup analyses, in which a significantly higher rate of symptoms alleviation was observed in patients with SOC plus HCQ versus SOC in the subgroup of patients without receiving potential anti-SARS-COV-2 drugs (Hazard ratio, 8.83, 95%CI, 1.09 to 71.3). No significance was observed in other subgroups. SOC=standard-of-care; HCQ=hydroxychloroquine; CI=confidence interval. NE=not estimable.

Changes of CRP values and blood lymphocyte count

Comparing to SOC alone, the addition of HCQ on SOC led to more rapid normalization of elevated baseline CRP and recovery of baseline lymphocytopenia, although the overall improvement rate become similar within the 28-day (Figure 4_panel A, B). The declined value of CRP from baseline by last assessment was significantly greater in SOC plus HCQ group than in the SOC group (absolute change, 6.986 versus 2.723 milligram/liter, P=0.045) (Figure 5). Similarly, the elevation of blood lymphocyte count at last assessment from baseline was greater in SOC plus HCQ group than that in the SOC group (absolute change, 0.062 versus 0.008 ×10⁹/liter, P=0.547) (Figure 5). Comprehensive analysis for other prespecified secondary outcomes including the reduction of erythrocyte sedimentation rate, IL-6 or TNF-α was not available due to very limited data of these parameters on pre-specified visiting date.

• Download figure
• Open in new tab

Figure 4.

Kaplan–Meier curves of the time to normalization of elevated CRP and recovery of lymphopenia in SOC plus HCQ versus SOC in the intention-to-treat population post-hoc subgroup analysis. Shown in panel A are data for 28 and 20 patients with elevated baseline CRP who were assigned to SOC plus HCQ and to SOC, respectively. The median time to normalization of elevated CRP in the SOC plus HCQ group (8 days; 95%CI, 5 to 14) was shorter to that in the SOC group (14 days; 95%CI, 5 to 21), but without statistical significance (P=0.42). Shown in panel B are data for 22 and 16 patients with baseline lymphopenia who were assigned to SOC plus HCQ and to SOC, respectively. The median time to recovery of lymphopenia in the SOC plus HCQ group (15 days; 95%CI, 7 to not estimable) was similar to that in the SOC group (15 days; 95%CI, 10 to 15) (P=0.76). SOC=standard-of-care; HCQ=hydroxychloroquine; CI=confidence interval. NE=not estimable.

• Download figure
• Open in new tab

Figure 5.

Absolute changes of CRP and Lymphocyte count from baseline by last assessment in all patients. Shown is the bar chart for the absolute changes of CRP and blood lymphocyte count at last assessment comparing to baseline values in patients with SOC with and without HCQ. Data were expressed as mean with standard error of mean. The declined value of CRP from baseline by last assessment was significantly greater in SOC plus HCQ group than in the SOC group (change, −6.986 versus −2.723 milligram/liter, P=0.045). The elevation of blood lymphocyte count at last assessment from baseline was greater in SOC plus HCQ group than that in the SOC group (change, 0.062 versus 0.008 ×10⁹/liter), but without statistical significance (P=0.547). Not all the patients have both baseline and last assessment for the calculation of value change in this analysis. Available number of patients in each group was provided in the bracket. SOC=standard-of-care; HCQ=hydroxychloroquine.

DISCUSSION

The present study (conducted during the outbreak of COVID-19 in China) is the first randomized, controlled trial showing that there is no increase of negative conversion rate of SARS-CoV-2 conferred by the addition of HCQ administration to the current SOC in patients hospitalized with COVID-19, including those received HCQ within or beyond 7 days of symptoms onset and those with or without receiving antiviral agents. An evident efficacy of HCQ on the alleviation of symptoms was demonstrated (Hazard ratio, 8.83, 95%CI, 1.09 to 71.3) in the subgroup of patients without receiving antiviral treatment in the post-hoc analysis. This was further supported by a significantly greater reduction of CRP (6.986 in SOC plus HCQ versus 2.723 in SOC, milligram/liter, P=0.045) conferred by the addition of HCQ, which also led to more rapid recovery of lymphopenia, albeit no statistical significance.

Our negative results on the anti-viral efficacy of HCQ obtained in this trial are on the contrary to the encouraging in-vitro results^4,5 and to the recently reported promising results from a non-randomized trial with 36 COVID-19 patients.⁶ Before interpreting our results, it should be noted that the precise anti-viral mechanism of CQ/HCQ has not been established in SARS-CoV-2, nor in other viruses.¹³ The discrepancy of the results between our trial and in-vitro studies highlighted the importance of pre-clinical in-vivo studies in which, the pharmacokinetic, pharmacodynamic and toxicity profiles of the tested drug should be established. Although there is no data of CQ/HCQ against SARS-CoV-2 from animal studies, data from a mice study suggested limited efficacy of CQ in inhibiting SARS-CoV replication.¹⁴ Promising anti-viral results of HCQ from the recent trial should be interpreted with caution due to its limited sample size and lack of randomization design. A combination therapy, HCQ plus azithromycin was recommended by the authors based on a 100% of virus negative conversion rate within one week of treatment in 6 patients with COVID-19.¹⁵. We were unable to validate this result due to the limited number of patients using azithromycin (N=2). In contrast to their excellent effects, our results did not support the clinical use of HCQ to suppress viral replication. One might argue that there was a median 16-day delay of HCQ treatment from symptoms onset in our study, which is longer than the time window of effective treatment of virus-related respiratory disease, like influenza.¹⁶ However, in the subgroup of patients who received HCQ within 7-day of illness onset, no additional virus negative conversion rate was observed beyond SOC in our trial. It is challenging to conduct a trial in hospitalized patients within 48-h of illness onset due to an estimated 12.5 day delay from illness onset to hospital admission.¹⁷ Future trials should consider enrolling patients at early-stage from the outpatient clinic. One might also argue that the current SOC included antivirals that were sources of confounding effects, although none of them has yet been proven efficacy in treating COVID-19. Nevertheless, no difference regarding time to virus negative conversion between the treatment and control groups was observed after excluding patients who ever received antivirals during the study period, including lopinavir-ritonavir, arbidol, oseltamivir, virazole, entecavir, ganciclovir and/or interferon-alpha. Of note, the promising drug, lopinavir-ritonavir which is an inhibitor of HIV protease, was recently demonstrated ineffective in lowering virus load of SARS-CoV-2.³ Moreover, the specimens collected in our trial for virus RNA determination was mostly from the upper respiratory tract, which introduced false-negative results.¹⁸ Bronchoalveolar lavage fluid, the best specimen for detection of SARS-CoV-2 was not well collected since the majority of our patients were mild/moderate cases without need for a bronchoscopy. But the prespecified definition for virus negative conversion was two negatives with at least 24 hours apart, which could significantly reduce false negativity. Although we did not monitor the concentration of HCQ in our study, the dosage we choose is enough to reach the 50% effective concentrations (EC₅₀) of HCQ against SARS-CoV-2.¹⁶ Therefore, it is not likely to have additional anti-viral effects by further escalating dosage. Taken together, future studies could take advantage of our results to design trials in more selective populations, at the earliest stage as possible (<48h of illness onset), and using more sensitive endpoints, such as viral load shedding. It also remains an open question whether HCQ would lead to an improved virologic response in patients with severe COVID-19.

In addition to virus infection, acute inflammation response is another hallmark of COVID-19.¹⁹ Recent findings in clinical series have shown that the systemic inflammation or cytokine storm is the driver of disease progression and death.^20,21 Substantially decrease of lymphocyte count and increase of inflammatory response marker, e.g., CRP were both observed in the early stage of patients who eventually progressed and died.²¹ These results highlighted the importance of the recovery of lymphopenia or anti-inflammation in preventing the development of systemic inflammation in critically ill COVID-19 patients. Such abilities and benefits were observed from HCQ in our current trial, showing that patients with SOC plus HCQ had a significantly greater reduction of CRP level and a moderate elevation of blood lymphocyte count at the last assessment comparing to patients with SOC only. These effects were observed after 5-day of HCQ treatment and maintained until the withdraw of HCQ. These encouraging results suggested clinical benefits of adding HCQ into the current standard management to limit inflammatory response, which is the key to prevent systemic inflammation and subsequent multiple organ failure and death. From mechanism perspective, HCQ as a weak base, accumulates within acidic vesicles, such as the lysosomes autophagosomes of phagocytic cells and changing local pH value, subsequently inhibits immune activation.²² Multiple cellular functions and molecular pathways were partially inhibited, such as MHC class II expression, antigen presentation or Toll-like receptor-7, −9 signaling pathways and production of IL-1, IL-6, TNF, and IFNγ²². These shreds of evidence support the anti-inflammatory effects of HCQ observed in our study, in which the patients were given HCQ with a loading dose of 1, 200 mg daily for three days followed by a maintained dose of 800 mg daily for remaining days (total treatment duration: 2 weeks or 3 weeks for mild/moderate or severe patients, respectively). Lower doses and shorter treatment duration are not recommended because a previous study using 400 mg of HCQ per day for 5 days did not show any additional effects on alleviating symptoms and viral suppression.²³ However, it remains to be determined whether the extension of treatment duration would bring more benefits because that HCQ has a gradual onset of action that might take weeks to reach maximal activity.²² It is also important to explore the combination therapy of HCQ with other anti-inflammatory drugs or higher dosage of HCQ in the treatment of patients with COVID-19.

Another encouraging result from our study is that using our current HCQ regimen, serious side-effects were rare and patients were well tolerated. This is in consistence with a previous report of cancer study using HCQ with a dose up to 1,200 mg daily,²⁴ Transient blurred vision was reported in one patient in our trial, of which, retinal damage was not assessed using the automated visual fields or spectral-domain optical coherence tomography as recommended for screening of retinopathy.²⁵ Although the retinal toxic effects from HCQ with a dosage of ≤5.0 mg/kg are less than 2% up to 10 years,²⁵ earlier development of retinal damage with a daily dose of 800 to 1,200mg was detected using sensitive retinal screening tests.²⁶ Therefore, the retinal damage could be underestimated in our trial. Evidence for increased rate of adverse gastrointestinal events in a high dose of HCQ²⁷ was observed in our trial, particularly an increased burden of diarrhea. Events of cardiac arrhythmia, e.g., prolonged QT interval²⁸ was not observed in our trial, possibly due to the relatively mild/moderate patients investigated or the short-term period of follow-up. However, with the increasing interest of the combined use of HCQ and azithromycin worldwide, physicians should be cautious of the increased risk of QT interval prolongation and fatal ventricular arrhythmia with azithromycin and other anti-microbials.^29,30 Overall, the generally favorable safety profiles of HCQ using our current regimen supported the future exploration of clinical benefits provided by this drug. However, drug-drug interaction¹⁰ should be taken into consideration when assessing safety and efficacy endpoints. The effects of HCQ in causing increased levels of digitoxin and metoprolol²² would be particularly relevant in severe COVID-19 patients and therefore would require close monitoring.

Numerous challenges were encountered in the performance of this multi-center randomized controlled trial, which was initiated during the most challenging time of the COVID-19 outbreak in China. Although a major strength of our trial was its randomized design, an open-label, as opposed to double-blind design, would introduce biased investigator-determined assessments. The knowledge on the association of clinical or virologic characteristics to disease course at the time of our trial design was largely limited. Selecting the virus negative conversion as the primary end-point might not be the most appropriate outcome for investigating the treatment efficacy of HCQ. The shortened time to alleviation of symptoms and rapid normalization elevated CRP and recovery of lymphocytopenia in patients who received HCQ observed in our trial suggested more potential of this drug in controlling acute inflammation and might be useful for preventing disease progression. Another difficulty is to ensure the fidelity to the protocol by investigators under highly challenging circumstances at the front lines in the COVID-19 treatment centers. A CRO was therefore hired to fully support the conduct and oversight of the trial. Population quarantine of Wuhan and neighboring cities, nation-wide travel restrictions and case/contact isolation were also barriers to collect and transfer data and paper files. The prespecified secondary endpoint of imaging changes on chest CT was therefore not finished by the cutoff date of analysis. The recruitment of eligible patients was unexpectedly difficult because that almost hundreds of clinical trials launched in the same period in response to the urgent call for the exploration of effective treatment against COVID-19 by the national health authorities. The rapid decline in eligible new cases of COVID-19 in the mid of March 2020 in China precluded the recruitment of our targeted number of cases. After a two-round extensive review of the efficacy and safety data generated from the interim analysis, the IDMC endorsed an early termination of the trial. Members from the IDMC all agreed that the conclusion drew from the trial is important for the proper use of HCQ in the clinical management of COVID-19, especially in overwhelming areas. The report of the trial could also be an important resource to facilitate a better design of future trials.

In conclusion, the results of our trial did not show additional virologic response by adding HCQ to the current SOC. The administration of HCQ with a loading dose of 1, 200 mg daily for three days followed by a maintained dose of 800 mg daily for remaining days (total treatment duration: 2 weeks or 3 weeks for mild/moderate or severe patients, respectively) was clearly associated with a greater decline of CRP, recovery of lymphopenia and a higher rate of symptoms alleviation and did not result in apparent safety concerns in patients hospitalized with COVID-19. As a well-tolerated, cheap, and widely available drug, future trials to determine the clinical benefits of HCQ in preventing disease progression is critically important considering the ongoing pandemic COVID-19.

Funding

This work was supported by the Emergent Projects of National Science and Technology (2020YFC0844500), National Natural Science Foundation of China (81970020, 81770025), National Key Research and Development Program of China (2016YFC0901104), Shanghai Municipal Key Clinical Specialty (shslczdzk02202, shslczdzk01103), National Innovative Research Team of High-level Local Universities in Shanghai, Shanghai Key Discipline for Respiratory Diseases (2017ZZ02014), National Major Scientific and Technological Special Project for Significant New Drugs Development (2017ZX09304007), Key Projects in the National Science and Technology Pillar Program during the Thirteenth Five-year Plan Period (2018ZX09206005-004, 2017ZX10202202-005-004, 2017ZX10203201-008). The funders played no role in study design, data collection, data analysis, data interpretation, or reporting. The guarantors had full access to all the data in the study, take responsibility for the integrity of the data and the accuracy of the data analysis, and had final responsibility for the decision to submit for publication.

Competing interests

All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

Ethical approval

The trial was approved by the Institutional Review Board of Ruijin Hospital, Shanghai Jiao Tong University School of Medicine (KY2020-29).

Patient consent

Written informed consent was obtained.

Data sharing

Anonymized datasets can be made available on reasonable request after approval from the trial management committee and after signing a data access agreement. Proposals should be directed to the corresponding author.

Transparency declaration

The lead author (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.

Dissemination to participants and related patient and public communities

This was a randomized controlled trial with no involvement of patients in the trial design, outcome measures, data analysis, results interpretation or manuscript writing. Personal health information used in this trial is not accessible to patients or the public. The study protocol is available online as a supplementary material with the publication of the paper. A preprint version of the study is publicly available on medRxiv.