BMJ

Confidential: For Review Only Hydroxychloroquine in patients mainly with mild to moderate COVID–19: an open–label, randomized, controlled trial

Journal: BMJ

Manuscript ID BMJ-2020-056844.R1

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the 22-Apr-2020 Author:

Complete List of Authors: Tang, Wei; Department of Pulmonary and Critical Care Medicine, Ruijin Hospital, Jiao Tong University School of Medicine, Shanghai, Cao, Zhujun; Department of Infectious Disease, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Han, Mingfeng; Department of Respiratory Medicine. No.2 people's hospital of Fuyang city, Fuyang, Anhui, China Wang, Zhengyan; Department of Respiratory Medicine, Hospital, University of Medicine, Suizhou, Hubei, China Chen, Junwen; No.1 People’s Hospital Affiliated hospital of Hubei University of Medicine Sun, Wenjin; Department of Infectious Diseases, Central Hospital of , Ezhou, Hubei, China Wu, Yaojie; Department of Cardiovascular Medicine, Yunmeng people's hospital, Xiaogan, Hubei, China Xiao, Wei; The First People’s Hospital of . Liu, Shengyong; Department of Infectious Diseases, Xiaogan hospital affiliated to university of science and technology, Xiaogan, Hubei, China Chen, Erzhen; Department of Emergency Medicine, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Chen, Wei; Department of Pulmonary and Critical Care Medicine, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China Wang, Xiongbiao; Department of respiratory medicine, Putuo Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, China Yang, Jiuyong; Department of Respiratory Medicine, Hubei Space Hospital of Xiaogan, Xiaogan, Hubei, China Lin, Jun; Department of Gastroenterology, Zhongnan Hospital of , Wuhan, Hubei, China Zhao, Qingxia; The Sixth People’s Hospital of , Department of Infectious Disease Yan, Youqin; Department of Infectious Disease, Wuhan No.7 Hospital, Wuhan, Hubei, China Xie, Zhibin; Departments of Respiratory Medicine, Xiaogan Hospital Affiliated to Wuhan University of Science and Technology, Xiaogan,

https://mc.manuscriptcentral.com/bmj Page 1 of 39 BMJ

1 2 3 Hubei, China 4 Li, Dan; Department of Respiratory Medicine, The Third People’s Hospital 5 of , Yichang, Hubei, China 6 Yang, Yaofeng; Department of Respiratory Medicine, Xiao Gan First 7 People’s Hospital, Xiaogan, Hubei province, China 8 Liu, Leshan; Clinical Research Center, Ruijin Hospital, Shanghai Jiao 9 Tong University School of Medicine, Shanghai, China Qu, Jieming; Rui Jin Hospital, School of Medicine, Shanghai Jiaotong 10 Confidential:University, Department For ofReview Respiratory Medicine Only 11 Ning, Guang; Shanghai Jiaotong University School of Medicine 12 Shi, Guochao; Department of Pulmonary and Critical Care Medicine, 13 Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, 14 Shanghai, China 15 Xie, Qing; Department of Infectious Disease, Ruijin Hospital, Shanghai 16 Jiao Tong University School of Medicine, Shanghai, China 17 Coronavirus disease, hydroxychloroquine, clinical oucome, treatment, Keywords: 18 novel coronavirus, SARS-CoV-2, negative conversion, adverse event 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 39

1 2 3 1 TITLE PAGE 4 5 2 Title: Hydroxychloroquine in patients mainly with mild to moderate COVID–19: an open–label, 6 7 8 3 randomized, controlled trial 9 10 4 Running title: Hydroxychloroquine for COVID–19 11 12 5 Wei Tang* Confidential:Email: [email protected], Forjob title: Reviewchest physician; associate Only professor 13 14 6 Zhujun Cao* Email: [email protected], job title: infectious diseases physician 15 16 7 Mingfeng Han* Email: [email protected], job title: chest physician 17 8 Zhengyan Wang* Email: [email protected], job title: chest physician 18 19 9 Junwen Chen Email: [email protected], job title: chest physician 20 21 10 Wenjin Sun Email: [email protected], job title: infectious diseases physician 22 23 11 Yaojie Wu Email: [email protected], job title: cardiovascular physician 24 25 12 Wei Xiao Email: [email protected], job title: chest physician 26 27 13 Shengyong Liu Email: [email protected], job title: infectious diseases physician 28 29 14 Erzhen Chen Email: [email protected], job title: professor 30 31 15 Wei Chen Email: [email protected], job title: chest physician 32 33 16 Xiongbiao Wang Email: [email protected], job title: chest physician 34 35 17 Jiuyong Yang Email: [email protected], job title: chest physician 36 18 Jun Lin Email: [email protected], job title: gastrointestinal physician 37 38 19 Qingxia Zhao E–mail: [email protected], job title: infectious diseases physician 39 40 20 Youqin Yan E–mail: [email protected], job title: infectious diseases physician 41 42 21 Zhibin Xie Email: [email protected], job title: chest physician 43 44 22 Dan Li Email: [email protected], job title: chest physician 45 46 23 Yaofeng Yang Email: [email protected], job title: chest physician 47 48 24 Leshan Liu E–mail: [email protected], job title: associate research fellow 49 50 25 Jieming Qu† Email: [email protected], job title: chest physician; professor 51 † 52 26 Guang Ning Email: [email protected], job title: endocrinology physician; professor 53 † 54 27 Guochao Shi Email: [email protected], job title: chest physician; professor 55 † 56 28 Qing Xie E–mail: [email protected], job title: professor 57 29 *Contributed equally 58 59 30 †Joint corresponding authors 60 1

https://mc.manuscriptcentral.com/bmj Page 3 of 39 BMJ

1 2 3 1 Affiliations: 4 5 6 2 1. Department of Pulmonary and Critical Care Medicine, Ruijin Hospital, Shanghai Jiao Tong 7 8 3 University School of Medicine, Shanghai, China (W Tang, W Chen, J Qu, G Shi) 9 10 11 4 2. Institute of Respiratory Diseases, School of Medicine, Shanghai Jiao Tong University, Shanghai, 12 Confidential: For Review Only 13 14 5 China (W Tang, W Chen, J Qu, G Shi) 15 16 6 3. Department of Infectious Disease, Ruijin Hospital, Shanghai Jiao Tong University School of 17 18 19 7 Medicine, Shanghai, China (Z Cao, Q Xie) 20 21 22 8 4. Department of Respiratory Medicine. No.2 people's hospital of Fuyang city, Fuyang, Anhui, 23 24 25 9 China (M Han) 26 27 10 5. Department of Respiratory Medicine, Suizhou Hospital, Hubei University of Medicine, Suizhou, 28 29 30 11 Hubei, China (Z Wang) 31 32 33 12 6. Department of Respiratory and Critical Care Medicine, Xiangyang No.1 People's Hospital, Hubei 34 35 13 University of Medicine, Xiangyang, Hubei, China (J Chen) 36 37 38 14 7. Department of Infectious Diseases, Central Hospital of Ezhou, Ezhou, Hubei, China (W Sun) 39 40 41 15 8. Department of Cardiovascular Medicine, Yunmeng people's hospital, Xiaogan, Hubei, China (Y 42 43 44 16 Wu) 45 46 17 9. Department of Respiratory Medicine, The First People's Hospital of Jingzhou City, Jingzhou, 47 48 49 18 Hubei, China (W Xiao) 50 51 52 19 10. Department of Infectious Diseases, Xiaogan hospital affiliated to Wuhan university of science 53 54 20 and technology, Xiaogan, Hubei, China (S Liu) 55 56 57 21 11. Department of Emergency Medicine, Ruijin Hospital, Shanghai Jiao Tong University School of 58 59 60 22 Medicine, Shanghai, China (E Chen) 2

https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 39

1 2 3 1 12. Department of Respiratory Medicine, Putuo Hospital, Shanghai University of Traditional 4 5 6 2 Chinese Medicine, Shanghai, China (X Wang) 7 8 3 13. Department of Respiratory Medicine, Hubei Space Hospital of Xiaogan, Xiaogan, Hubei, China 9 10 11 4 (J Yang) 12 Confidential: For Review Only 13 14 5 14. Department of Gastroenterology, Zhongnan Hospital of Wuhan University, Wuhan, Hubei, 15 16 6 China (J Lin) 17 18 19 7 15. Department of Infectious Disease, The Sixth People’s Hospital of Zhengzhou, Zhengzhou, 20 21 22 8 Henan, China (Q Zhao) 23 24 25 9 16. Department of Infectious Disease, Wuhan No.7 Hospital, Wuhan, Hubei, China (Y Yan) 26 27 10 17. Departments of Respiratory Medicine, Xiaogan Hospital Affiliated to Wuhan University of 28 29 30 11 Science and Technology, Xiaogan, Hubei, China (Z Xie) 31 32 33 12 18. Department of Respiratory Medicine, The Third People’s Hospital of Yichang, Yichang, Hubei, 34 35 13 China (D Li) 36 37 38 14 19. Department of Respiratory Medicine, Xiao Gan First People’s Hospital, Xiaogan, Hubei 39 40 41 15 province, China (Y Yang) 42 43 44 16 20. Clinical Research Center, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, 45 46 17 Shanghai, China (L Liu) 47 48 49 18 21. Shanghai National Research Centre for Endocrine and Metabolic Diseases, State Key Laboratory 50 51 52 19 of Medical Genomics, Shanghai Institute for Endocrine and Metabolic Diseases, Ruijin Hospital, 53 54 20 Shanghai Jiao Tong University School of Medicine, Shanghai, China (G Nin) 55 56 57 58 59 60 3

https://mc.manuscriptcentral.com/bmj Page 5 of 39 BMJ

1 2 3 1 Correspondence: 4 5 6 2 Qing Xie, 7 8 3 Department of Infectious Disease, Ruijin Hospital, 9 10 11 4 Shanghai Jiao Tong University School of Medicine, 12 Confidential: For Review Only 13 14 5 197 Ruijin 2nd Road, Shanghai, 200025, China 15 16 6 Email: [email protected] 17 18 19 7 Author contributions: 20 21 22 8 WT, ZC, MH and ZW contributed equally to this paper and share joint first authorship. QX, GS, GN, 23 24 25 9 and JQ contributed equally to this paper and share joint corresponding authorship. QX and GS were 26 27 10 co–chief investigators of this trial and act as guarantors of the study in its entirety. QX, GS, GN, and 28 29 30 11 JQ led the development of the research question, study design, and obtaining the funding. WT, GS, 31 32 33 12 QX, GN, JQ, and LL coordinated the operational delivery of the study protocol to the coordinating 34 35 13 centers. MH, ZW, JC, WS, YW, WX, SL, EC, WC, XW, JY, JL, QZ, YY, ZX, DL, and YY 36 37 38 14 represented the collaborating coordinating centers responsible for their centers’ participation in the 39 40 41 15 trial. ZC led and produced the first draft of this manuscript. All authors provided critical review and 42 43 44 16 final approval of the manuscript. The corresponding author attests that all listed authors meet 45 46 17 authorship criteria and that no others meeting the criteria have been omitted. 47 48 49 50 51 52 53 54 55 56 57 58 59 60 4

https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 39

1 2 3 1 What is already known on this topic 4 5 6 2 —The pandemic of coronavirus disease 2019 (COVID–19) imposes substantial burdens on 7 8 3 individuals, communities, health–care facilities, markets, governments, etc. globally. 9 10 11 4 —There is no specific treatment approved for COVID–19 or vaccine to prevent infection with the 12 Confidential: For Review Only 13 14 5 novel coronavirus. 15 16 6 —During the urgent pandemic, media headlines the utility of drugs without solid evidence but buries 17 18 19 7 the side–effects of these drugs. 20 21 22 8 What this study adds 23 24 25 9 —In this randomized clinical trial of patients mainly with persistent mild to moderate COVID–19, 26 27 10 exposure to hydroxychloroquine led to similar rates of virus elimination comparing to the current 28 29 30 11 standard–of–care. 31 32 33 12 —Adverse events, mostly gastrointestinal related, were significantly increased in patients who 34 35 13 received hydroxychloroquine. 36 37 38 14 —Overall, the results from our trial do not support the use of hydroxychloroquine in patients with 39 40 41 15 persistent mild to moderate COVID–19. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 5

https://mc.manuscriptcentral.com/bmj Page 7 of 39 BMJ

1 2 3 1 Abstract 4 5 2 Objectives To assess the efficacy and safety of hydroxychloroquine (HCQ) plus standard–of–care 6 7 8 3 (SOC) compared with SOC alone in adult patients with COVID–19. 9 10 11 4 Design Multicenter, open–label, randomized controlled trial. 12 Confidential: For Review Only 13 5 Setting 16 government–designated COVID–19 treatment centers in China through 11 to 29 in 14 15 16 6 February 2020. 17 18 7 Participants 150 patients hospitalized with laboratory confirmed COVID–19 were included in the 19 20 21 8 intention to treat analysis. 75 patients were assigned to HCQ plus SOC and 75 to SOC alone. 22 23 24 9 Interventions HCQ was administrated with a loading dose of 1, 200 mg daily for three days followed 25 26 10 by a maintained dose of 800 mg daily for the remaining days (total treatment duration: 2 or 3 weeks 27 28 29 11 for mild/moderate or severe patients, respectively). 30 31 12 Main outcome measures The primary outcome was overall 28–day negative conversion rate of 32 33 34 13 SARS–CoV–2 by intention to treat analysis. Adverse events were analyzed in the safety population. 35 36 37 14 Results Among 150 patients, 148 were with mild to moderate disease and 2 were with severe disease. 38 39 15 The median time from randomization to illness onset was 16.6±10.5 days. Overall, the negative 40 41 42 16 conversion rate within 28–day in SOC plus HCQ group was 85.4% (95% confidence interval (CI) 43 44 17 73.8% to 93.8%), similar to that in the SOC group 81.3% (95%CI 71.2% to 89.6%). Between-group 45 46 47 18 difference was 4.1% (95%CI –10.3% to18.5%). Adverse events were reported in 7 (8.8%) patients 48 49 50 19 without receiving HCQ (N=80) and in 21 (30%) HCQ recipients (N=70) with two serious adverse 51 52 20 events. The most common adverse event in the HCQ recipients was diarrhea, reported in 7 (10%) 53 54 55 21 patients. 56 57 22 Conclusions The administration of HCQ did not result in a significantly higher negative conversion 58 59 60 23 rate than SOC alone in patients mainly hospitalized with persistent mild to moderate COVID–19. 6

https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 39

1 2 3 1 Adverse events were significantly increased in HCQ recipients. 4 5 2 Trial registration ChiCTR2000029868. 6 7 8 3 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 7

https://mc.manuscriptcentral.com/bmj Page 9 of 39 BMJ

1 2 3 1 Print abstract 4 5 2 Study question To assess the efficacy and safety of hydroxychloroquine (HCQ) plus standard–of– 6 7 8 3 care (SOC) compared with SOC alone in adult patients with COVID–19. 9 10 11 4 Methods This is a multicenter, open–label, randomized controlled trial conducted in 16 government– 12 Confidential: For Review Only 13 5 designated COVID–19 treatment centers in China through 11 to 29 in February 2020. A total of150 14 15 16 6 patients hospitalized with laboratory confirmed COVID–19 were included in the intention to treat 17 18 7 analysis. Among them, 75 patients were assigned to HCQ plus SOC and 75 to SOC alone. HCQ was 19 20 21 8 administrated with a loading dose of 1, 200 mg daily for three days followed by a maintained dose of 22 23 24 9 800 mg daily for the remaining days (total treatment duration: 2 or 3 weeks for mild/moderate or 25 26 10 severe patients, respectively). The primary outcome was overall 28–day negative conversion rate of 27 28 29 11 SARS–CoV–2 by intention to treat. Adverse events were analyzed in the safety population. 30 31 12 Study answer and limitations Among 150 patients, 148 were with mild to moderate disease, and 2 32 33 34 13 were with severe disease. The median time from randomization to illness onset was 16.6±10.5 days. 35 36 37 14 Overall, the negative conversion rate within 28–day in SOC plus HCQ group was 85.4% (95% 38 39 15 confidence interval (CI) 73.8% to 93.8%), similar to that in the SOC group 81.3% (95%CI 71.2% to 40 41 42 16 89.6%). Between-group difference was 4.1% (95%CI –10.3% to18.5%). Adverse events were 43 44 17 reported in 7 (8.8%) patients without receiving HCQ and in 21 (30%) HCQ recipients with two 45 46 47 18 serious adverse events. The most common adverse event in the HCQ recipients was diarrhea, reported 48 49 50 19 in 7 (10%) patients. The major limitations of the study are its early termination with underpowered 51 52 20 sample size and its open-label design as opposed to double blind. 53 54 55 21 What this study adds Our trial does not support the use of hydroxychloroquine in patients with 56 57 22 persistent mild to moderate COVID–19 due to limited effects on virus eliminating and significantly 58 59 60 23 increased adverse events. 8

https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 39

1 2 3 1 Funding, competing interests, data sharing This work was supported by the Emergent Projects of 4 5 2 National Science and Technology (2020YFC0844500), National Natural Science Foundation of 6 7 8 3 China (81970020, 81770025), National Key Research and Development Program of China 9 10 11 4 (2016YFC0901104), Shanghai Municipal Key Clinical Specialty (shslczdzk02202, shslczdzk01103), 12 Confidential: For Review Only 13 5 National Innovative Research Team of High–level Local Universities in Shanghai, Shanghai Key 14 15 16 6 Discipline for Respiratory Diseases (2017ZZ02014), National Major Scientific and Technological 17 18 7 Special Project for Significant New Drugs Development (2017ZX09304007), Key Projects in the 19 20 21 8 National Science and Technology Pillar Program during the Thirteenth Five–year Plan Period 22 23 24 9 (2018ZX09206005–004, 2017ZX10202202–005–004, 2017ZX10203201–008). All authors declared 25 26 10 no competing interests. Anonymized datasets can be made available on reasonable request after 27 28 29 11 approval from the trial management committee. 30 31 12 Study registration ChiCTR2000029868 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 9

https://mc.manuscriptcentral.com/bmj Page 11 of 39 BMJ

1 2 3 1 INTRODUCTION 4 5 2 The coronavirus disease 2019 (COVID–19) caused by the severe acute respiratory syndrome 6 7 8 3 coronavirus 2 (SARS–CoV–2) has swept into 185 countries/regions within five months. As of 22 9 10 1 11 4 April, more than 2.5 million infections and 178 thousand deaths have been reported. 12 Confidential: For Review Only 13 5 Several agents or drugs including, remdesivir, favipiravir, ribavirin, lopinavir–ritonavir (used in 14 15 16 6 combination) and chloroquine (CQ) or hydroxychloroquine (HCQ), have been highlighted based on 17 18 7 the promising in–vitro results and therapeutic experiences from another two coronavirus diseases 19 20 21 8 including the severe acute respiratory syndrome and the Middle East respiratory syndrome.2 However, 22 23 24 9 none of these promising results has yet been translated into clinical benefits of patients with COVID– 25 26 10 19, including lopinavir–ritonavir, reported from the most recently failed trial.3 27 28 29 11 CQ and its hydroxy–analog HCQ, best known as antimalarial drugs, are glaring on the list of COVID– 30 31 12 19 therapy, due to potent antiviral activity against SARS–CoV–2 from in–vitro studies,4,5 and 32 33 34 13 promising results from news reports of some ongoing trials.6 Despite their unclear benefits, CQ and 35 36 37 14 HCQ are both recommended for off–label use in the treatment of COVID–19 by the Chinese National 38 39 15 guideline7 and recently authorized by the U.S. Food and Drug Administration for emergency use.8 40 41 42 16 HCQ was also recently recommended by the American president Donald Trump. Such a presidential 43 44 17 endorsement stimulates an avalanche of demand for HCQ, which buried the dark–side of this drug. 45 46 47 18 Deaths have been reported in Nigeria among people self–treating for apparent COVID–19 with CQ 48 49 9 50 19 overdoses. Retinopathy, gastrointestinal, and cardiac side effects are well documented with the use 51 52 20 of CQ or HCQ in the treatment of malarial and rheumatic diseases.10 HCQ is preferred in clinical 53 54 10 55 21 applications due to its lower toxicity, particularly retinal toxicity, and three times the potency 56 57 22 against SARS–CoV–2 infection comparing to CQ in the recent in–vitro study.5 Currently, there is no 58 59 60 23 convincing evidence from well–designed clinical trials to support the use of CQ/HCQ with good 10

https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 39

1 2 3 1 efficacy and safety for the treatment of COVID–19. Rapidly conduction of such trails with high– 4 5 2 quality is challenging in the face of a dangerous coronavirus outbreak, in which, healthcare workers 6 7 8 3 are under overwhelming work and highest risk of exposure to developing COVID–19.11 9 10 11 4 Having encountered numerous challenges, we conducted a multicenter, open–label, randomized, 12 Confidential: For Review Only 13 5 controlled trial to assess the efficacy and safety of HCQ sulfate in adult patients with COVID–19. A 14 15 16 6 clearer verdict will come from such a trial for the use of HCQ in patients with COVID–19. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 11

https://mc.manuscriptcentral.com/bmj Page 13 of 39 BMJ

1 2 3 1 METHODS 4 5 6 2 Trial oversight 7 8 3 The study was designed and initiated by the principal investigators after the protocol was approved 9 10 11 4 by the institutional review board in Ruijin Hospital on February 6, 2020. The protocol and approval 12 Confidential: For Review Only 13 14 5 documents are available online with the full text of this article. Written informed consent was 15 16 6 obtained from all patients. Hospitals with the capability of providing the current standard–of–care 17 18 19 7 (SOC) for COVID–19 were invited to participate in the study by the principal investigators. Minimum 20 21 22 8 requirements for the SOC included the provision of intravenous fluids, supplemental oxygen, regular 23 24 25 9 laboratory testing, and SARS–CoV–2 test, hemodynamic monitoring, and intensive care, and the 26 27 10 ability to deliver concomitant medications. The trial was conducted urgently during the outbreak of 28 29 30 11 COVID–19 in compliance with the Declaration of Helsinki, Good Clinical Practice guidelines, and 31 32 33 12 local regulatory requirements. The Shanghai Pharmaceuticals Holding Co.,Ltd donated the 34 35 13 investigated drug, HCQ, but was not involved in the study design, accrual, analyses of data, or 36 37 38 14 preparation of the manuscript. The principal investigators designed the study and made the decision 39 40 41 15 to submit the manuscript for publication All authors vouch for the veracity of the data, analyses, and 42 43 44 16 trial protocol and fidelity of the study to the protocol. A contract research organization, R&G 45 46 17 PharmaStudies Co., Ltd., was hired to conduct the study, collect data, and perform statistical analyses. 47 48 49 18 Data were recorded by clinical research coordinators followed by queries from clinical research 50 51 52 19 associates. Confirmed data were then entered into the Web–based OpenClinica database for statistical 53 54 20 analyses performed by the study statistician and reviewed by the senior statistician in accordance with 55 56 57 21 Good Clinical Practice guidelines. Data collection forms and statistical analysis plans are available 58 59 60 22 online with the full text of this article. 12

https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 39

1 2 3 1 An independent data and safety monitoring committee (IDMC) periodically reviewed the progress 4 5 6 2 and oversight of the study. The interim analysis was performed on March 14 and the results were 7 8 3 presented to the IDMC. The rapid decline in eligible new cases of COVID–19 at that 9 10 11 4 precluded the recruitment of our targeted number of cases. After a two–round extensive review of the 12 Confidential: For Review Only 13 14 5 efficacy and safety data generated from the interim analysis, the IDMC endorsed an early termination 15 16 6 of the trial. Members from the committee all agreed that the data from the trial is important for 17 18 19 7 clinicians, the public, and the government to avoid inappropriate use of HCQ in the clinical 20 21 22 8 management of COVID–19, particularly in overwhelming areas. The report of the trial could 23 24 25 9 potentially be an important resource to facilitate a better design of future trials. 26 27 10 Trail design, Randomization, and procedures 28 29 30 11 This study was a multicenter, randomized, parallel, open–label, trial of HCQ in hospitalized patients 31 32 33 12 with COVID–19. Patients were enrolled by the site investigators in 16 government–designated 34 35 13 COVID–19 treatment centers from three provinces in China (Hubei, Henan and Anhui province). No 36 37 38 14 placebo was used and drugs were not masked. Stratified random sampling was applied to stratify all 39 40 41 15 eligible patients according to the disease severity (mild to moderate or severe) followed by random 42 43 44 16 assignment (in a 1:1 ratio) in each stratum to ensure a balanced distribution of disease severity 45 46 17 between treatment (HCQ plus SOC) and control (SOC only) groups. Randomization was performed 47 48 49 18 by the statistician. Equal numbers of cards with each group assignment number randomly generated 50 51 52 19 by computer were placed in sequentially numbered envelopes that were opened as the patients were 53 54 20 enrolled. All patients were managed with SOC aligning with the indications from the updating 55 56 57 21 National clinical practice guidelines for COVID–19 in China. Patients in the treatment group were 58 59 60 22 administrated with HCQ within 24 hours after randomization and were administrated with a loading 13

https://mc.manuscriptcentral.com/bmj Page 15 of 39 BMJ

1 2 3 1 dose of 1, 200 mg daily for three days followed by a maintained dose of 800 mg daily for remaining 4 5 6 2 days (total treatment duration: 2 weeks or 3 weeks for mild/moderate or severe patients, respectively). 7 8 3 The dose for HCQ was adjusted when adverse events were related to HCQ as judged by investigators. 9 10 11 4 Details for dose adjustment were provided in the study protocol available online. Neither patients, 12 Confidential: For Review Only 13 14 5 nor investigators, nor statisticians were masked to treatment assignment. Lab technicians who 15 16 6 performed virologic, chemistry, and other routine measurement were unaware of treatment 17 18 19 7 information. 20 21 22 8 Patients 23 24 25 9 Inclusion criteria: 1) age >18 years; 2) with ongoing SARS–CoV–2 infection confirmed in upper or 26 27 10 lower respiratory tract specimens with real–time reverse–transcriptase–polymerase–chain–reaction 28 29 30 11 (RT–PCR); 3) willingness to participate; 4) consent not to be enrolled by other clinical trials during 31 32 33 12 the study period. Pneumonia on chest computed tomography was not mandatory for inclusion. 34 35 13 Exclusion criteria: 1) age <18 years; 2) with severe conditions including malignancies, 36 37 38 14 heart/liver/kidney disease or poorly controlled metabolic diseases; 3) not suitable to be administrated 39 40 41 15 through the gastrointestinal tract;4) pregnant or lactation; 5) allergy to HCQ; 6) unable to cooperate 42 43 44 16 with investigators due to cognitive impairments or poor mental status; 7) with severe liver disease 45 46 17 (e.g. Child Pugh grade C, ALT >5 fold of times upper limit);8) with severe renal impairment 47 48 49 18 (estimated glomerular filtration rate <30 mL/min/1.73 m2) or receiving continuous renal replacement 50 51 52 19 therapy, hemodialysis, peritoneal dialysis. In the original protocol, patients with severe COVID–19 53 54 20 were excluded. Considering the anti–inflammatory property of HCQ might favor disease regression, 55 56 57 21 we decided to include patients with severe COVID–19. (change approved by the Ethics Committee 58 59 60 22 on February 17, 2020). 14

https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 39

1 2 3 1 Definition for disease severity of COVID–19 was based on the 5th version of Chinese guideline for 4 5 12 6 2 the management of COVID–19 : Mild: patients with mild symptoms but no pneumonia 7 8 3 manifestation on imaging; Moderate: patients with fever, cough, sputum production, other respiratory 9 10 11 4 tract or non–specific symptoms along with pneumonia manifestation on imaging but no signs of 12 Confidential: For Review Only 13 14 5 severe pneumonia defined as the presence of SaO2/SPO2 <94% on room air or PaO2/FiO2 ratio 15 16 6 ≤300mgHg. 17 18 19 7 Assessment 20 21 22 8 Upper and/or lower respiratory tract specimens were obtained from each patient upon screening (Day 23 24 25 9 –3~1), during treatment and post–treatment follow–up at scheduled visits on days 4, 7, 10, 14, 21 and 26 27 10 28. SARS–CoV–2 was measured by the local Center for Disease Control and Prevention or authorized 28 29 30 11 health institutions or hospitals at each site using assays that have been approved by the National 31 32 33 12 Medical Products Administration. Measurements were performed based on the recommendations by 34 35 13 the National Institute for Viral Disease Control and Prevention (China) (http://ivdc.chinacdc.cn/kyjz/ 36 37 38 14 202001/t20200121_211337.html). Methods for total RNA extraction and amplification through RT– 39 40 41 15 PCR were similar as described elsewhere.13 Instead of quantitative data (cycle threshold value) 42 43 44 16 reported from RT–PCR assay, we only collected qualitative data reported from our trail sites. Based 45 46 17 on national recommendation, a cycle threshold value less than 37 was defined as a positive test result, 47 48 49 18 and a cycle threshold value of 40 or more was defined as a negative test. Cycle threshold values 50 51 52 19 between 37 and 40 confirmed by retesting were reported as unclassified. 53 54 20 In addition to SARS–CoV–2, patients were assessed on each scheduled visit for vital signs, C– 55 56 57 21 reactive protein (CRP), erythrocyte sedimentation rate (ESR), Tumor necrosis factor (TNF)–α, 58 59 60 22 Interleukin–6 (IL–6), complete blood cells count with differential, blood chemistry, coagulation panel, 15

https://mc.manuscriptcentral.com/bmj Page 17 of 39 BMJ

1 2 3 1 pulse oximetry, and respiratory symptoms. Administration records of HCQ and adverse events were 4 5 6 2 reviewed daily to ensure fidelity to the protocol and more importantly, patient safety. Chest computed 7 8 3 tomography was assessed upon screening and at the last visit of the treatment period (Day 14 for mild 9 10 11 4 to moderate and Day 21 for severe patients). Chest computed tomography examinations can be 12 Confidential: For Review Only 13 14 5 exempted by the investigators if the participants can provide qualified examination results within 3 15 16 6 days before the initiation of the study. More details for data collection were provided in the protocol 17 18 19 7 available online with full text of this article. 20 21 22 8 Endpoint 23 24 25 9 The primary endpoint for this trial was the negative conversion rate of SARS–CoV–2 within 28–day 26 27 10 for all patients and clinical improvement within 28–day only in severe patients. However, since the 28 29 30 11 trial was stopped early and only 2 patients with severe disease were enrolled, results on clinical 31 32 33 12 improvement are not presented. We focused on the report of the primary endpoint in this paper. 34 35 13 Negative conversion of SARS–CoV–2 was defined as two consecutive confirmation of "Negatives" 36 37 38 14 reported at least 24 hours apart without subsequent report of "Positive" SARS–CoV–2 by the end of 39 40 41 15 the study. The date of the first "Negative" report was considered as the date of negative conversion. 42 43 44 16 In the original protocol, the primary endpoint was prespecified as the “Negative conversion rate by 45 46 17 Day10”. We modified this to “Negative conversion rate within 28day” to test the overall difference 47 48 49 18 of negative conversion rate between the treatment and control groups. Negative conversion rate at 50 51 52 19 Day4, 7, 10, 14, or 21 was subsequently specified as a secondary endpoint as listed in the protocol 53 54 20 but they do not appear on the trial registration list. (approved on February 17, 2020, by the Ethics 55 56 57 21 Committee). The listed secondary endpoint in the trial registration was adverse events coded using 58 59 60 22 the latest version of Medical Dictionary for Regulatory Activities coding dictionary, recorded in 16

https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 39

1 2 3 1 standard medical terminology and graded according to the National Cancer Institute Common 4 5 6 2 Terminology Criteria for Adverse Events. 7 8 3 Other pre–specified secondary endpoints not listed in the trial registration but included in the protocol 9 10 11 4 were 1) alleviation of clinical symptoms; 2) improvement rate of CRP, ESR, IL–6, TNF–α and 12 Confidential: For Review Only 13 14 5 absolute blood lymphocyte count; 3) improvement rate of lung lesions on chest radiology; 4) all– 15 16 6 cause mortality; 5) disease progression rate in mild to moderate patients. The time frame for these 17 18 19 7 secondary endpoints were all from randomization to 28–day. Pre–specified secondary endpoints for 20 21 22 8 severe patients were not listed here but in the protocol. Due to the early termination of our study, we 23 24 25 9 could not able to justify the results from these analyses with an underpowered sample size and 26 27 10 therefore decided not to emphasize them in this article to avoid misinterpretation. The only one 28 29 30 11 presented here was the alleviation of clinical symptoms within 28 day which is an important outcome 31 32 33 12 of interest prespecified in our protocol. Definition for the alleviation of clinical symptoms was 1) 34 35 13 resolving from fever to an axillary temperature of ≤36.6℃ and; 2) normalization of SpO2 (>94% on 36 37 38 14 room air) and; 3) disappearance of respiratory symptoms including nasal congestion, cough, sore 39 40 41 15 throat, sputum production and shortness of breath. 42 43 44 16 In addition, changes of C–reactive protein and absolute blood lymphocyte count from baseline to last 45 46 17 follow–up were analyzed post–hoc. 47 48 49 18 Statistical analysis 50 51 52 19 The sample size was calculated based on the alternative hypothesis of a 30% increase in the speed of 53 54 20 virus nucleic acid negativity, therefore, a total of 248 events is needed with a Log–Rank test. An 55 56 57 21 interim analysis was planned when around 150 patients were treated for at least 7 days. O’Brien– 58 59 60 22 Fleming cumulative α–spending function by Lan–DeMets algorithm ( Lan–Demets, 1983 ) was 17

https://mc.manuscriptcentral.com/bmj Page 19 of 39 BMJ

1 2 3 1 applied to control family–wise type–I error. In the original protocol, the target number of enrollments 4 5 6 2 was set to 200 without a prior sample size calculation. We updated the trial design to enroll 7 8 3 approximately 360 subjects (180 per group) to assure a power of 80% and the family–wise (in the 9 10 11 4 entire trial) type–I error ≤0.05 (approved on February 10, 2020, by the Ethics Committee). 12 Confidential: For Review Only 13 14 5 The overall negative conversion rate was estimated by analyzing time to negative conversion of 15 16 6 SARS–CoV–2 using the Kaplan–Meier method in the intention–to–treat population and compared 17 18 19 7 with a log-rank test. Patients failed to reach the negative conversion of SARS–CoV–2 (see definition 20 21 22 8 in Endpoint section) by the cutoff date of the analysis (14 March 2020) was considered as right- 23 24 25 9 censored at the last visit date. All these patients remained in hospital and may reach negative 26 27 10 conversion after our last visit date, but the specific timing of the event is unknown. Similar approaches 28 29 30 11 were applied to the analysis of symptoms alleviation. Between-group rate difference was used to 31 32 33 12 show treatment effect size and 95% confidence interval (CI) was estimated by approximating normal 34 35 13 distribution and the standard deviation was estimated by the bootstrap method (N=1000). Two key laboratory 36 37 38 14 parameters, CRP, and blood lymphocyte were analyzed post hoc. Absolute changes from baseline of CRP 39 40 41 15 and blood lymphocyte count by last assessment were compared between actual treatment groups 42 43 44 16 using the Mann-Whitney-Wilcoxon test based on data distribution. Between-group mean difference 45 46 17 of change and their 95% were used to show treatment effect size. In addition, a final score for CRP 47 48 49 18 or blood lymphocyte count by adjusting the baseline score using the analysis of covariance 50 51 14 52 19 (ANCOVA) model to control for baseline imbalance. Safety analyses were based on the patients’ 53 54 20 actual treatment exposure. Data analyses were conducted on SAS version 9.4. 55 56 57 21 Patient involvement 58 59 60 22 No patients were involved in setting the research question or the outcome measures, nor were they 18

https://mc.manuscriptcentral.com/bmj BMJ Page 20 of 39

1 2 3 1 involved in developing plans for recruitment, design or implementation of the study. No patients were 4 5 6 2 asked to advise on interpretation or writing up of results. There are no plans to disseminate the results 7 8 3 of the research to study participants or the relevant patient community. 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 19

https://mc.manuscriptcentral.com/bmj Page 21 of 39 BMJ

1 2 3 1 RESULTS 4 5 2 Patient 6 7 8 3 A total of 191 patients admitted with COVID–19 from February 11, 2020, to February 29, 2020, were 9 10 11 4 assessed for eligibility, of which 41 did not meet eligibility criteria. The remaining 150 patients 12 Confidential: For Review Only 13 5 underwent randomization; Among them, 75 patients were assigned to SOC and 75 patients to SOC 14 15 16 6 plus HCQ group (Figure 1). The mean age of the patients was 46 years and 82 (55%) were male. The 17 18 7 mean day from disease onset to randomization was 16.6±10.5 days and 90 (60%) patients had 19 20 21 8 concomitant medication before randomization; among them, 52 (34.7%) received antiviral treatment 22 23 24 9 (Table 1). 148 patients (99%) had mild to moderate COVID–19 and only 2 patients (1%) were severe 25 26 10 upon screening. Baseline demographic, epidemiological, and clinical characteristics of the patients 27 28 29 11 between the two groups are shown in Table 1. 30 31 12 By 14 March 2020 (the cutoff date for data analysis), the median duration of follow–up was 21 days 32 33 34 13 (range, 2 to 33) in the SOC group and 20 days (range, 3 to 31) in the SOC plus HCQ group. Of the 35 36 37 14 75 patients assigned to receive SOC plus HCQ, 6 patients did not receive any dose of HCQ; of them, 38 39 15 3 patients withdrew consent and 3 patients refuse to be administrated HCQ. The co–intervention 40 41 42 16 including antiviral, antibiotics, and systemic glucocorticoid therapy between the two groups were 43 44 17 similar (Table 2). One patient with the moderate disease in the HCQ group progressed to severe 45 46 47 18 COVID–19 and no patients died during follow–up. 48 49 50 19 Primary Outcome 51 52 20 Overall, the negative conversion rate of SARS–CoV–2 among patients who were assigned to receive 53 54 55 21 SOC plus HCQ was 85.4% (95%CI 73.8% to 93.8%), similar to that of the SOC group 81.3% (95%CI 56 57 22 71.2% to 89.6%) within 28–day. The difference of overall negative conversion rate between SOC 58 59 60 23 plus HCQ and SOC alone was 4.1% (95%CI –10.3% to 18.5%). The median time to negative 20

https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 39

1 2 3 1 conversion was also similar in the SOC plus HCQ group (8 days, 95%CI 5 to 10 days) with that in 4 5 2 the SOC group (7 days, 95%CI 5 to 8 days) (Figure 2). 6 7 8 3 Safety 9 10 11 4 Six patients assigned to the SOC plus HCQ group but did not receive HCQ treatment were classified 12 Confidential: For Review Only 13 5 as HCQ non–recipient in the safety population. One patient in the SOC group wrongly received 14– 14 15 16 6 day of HCQ treatment with an accumulative dose of 11, 600 mg. This patient was classified as HCQ 17 18 7 recipient in the safety population (Figure 1). Safety endpoints were compared between HCQ recipient 19 20 21 8 and non–recipient (Table 3). In HCQ recipients, the median duration of HCQ treatment was 14 days 22 23 24 9 (range, 1 to 22). Between randomization and final visit, a total of 21 patients (30%) in the SOC plus 25 26 10 HCQ group reported adverse events, significantly (P=0.001) higher than those (7 patients, 8.8%) 27 28 29 11 reported in the SOC group (Table 3). No serious adverse events were reported in the SOC group. Two 30 31 12 patients in the HCQ group reported serious adverse events due to disease progression and upper 32 33 34 13 respiratory infection. The case with upper respiratory infection discharged after finishing the 14–day 35 36 37 14 treatment of HCQ and developed throat–drying and pharyngalgia requiring re-admission without 38 39 15 evidence of pneumonia on chest computed tomography during the extended follow–up period. 40 41 42 16 The most common adverse events in the SOC plus HCQ group were diarrhea, which was more 43 44 17 frequent than that in the SOC group (10% versus 0%, P=0.004). HCQ was discontinued in one patient 45 46 47 18 due to blurred vision and was adjusted to give a lower dose in one patient who reported thirst. These 48 49 50 19 two adverse events were both transient with a period of 1–2 day. 51 52 20 Secondary Outcomes 53 54 55 21 Negative conversion rate by a specific time–point, 4–, 7–, 10–, 14– or 21–day was also similar 56 57 22 between the two groups (Supplementary Table 1). The overall rate of symptoms alleviation within 58 59 60 23 28–day was similar between patients with SOC with (59.9%, 95%CI 45.0% to 75.3%) and without 21

https://mc.manuscriptcentral.com/bmj Page 23 of 39 BMJ

1 2 3 1 HCQ (66.6%, 95%CI 39.5% to 90.9%). Between-group difference was –6.6% (95%CI –41.3% to 4 5 2 28.0%). The median time to alleviation of clinical symptoms was similar in the SOC plus HCQ group 6 7 8 3 with that in the SOC group (19 days versus 21 days). 9 10 11 4 Post–hoc analysis 12 Confidential: For Review Only 13 5 The absolute change of CRP from baseline by last assessment was numerically greater in SOC plus 14 15 16 6 HCQ group than in the SOC group (–6.99±15.00 mg/L versus –2.72±9.03 mg/L, p=0.12 by Mann– 17 18 7 Whitney–Wilcoxon test). The mean difference of CRP change between the two groups (HCQ–SOC) 19 20 21 8 was –4.26 mg/L (95%CI –8.48 to –0.05 mg/L). The absolute change of blood lymphocyte count from 22 23 24 9 baseline by last assessment was numerically greater in SOC plus HCQ group than in the SOC group 25 26 10 (0.06±0.56×109/L versus 0.01±0.50×109/L, p=0.57 by Mann–Whitney–Wilcoxon test). The mean 27 28 9 29 11 difference of lymphocyte count change between the two groups (HCQ–SOC) was 0.05×10 /L 30 31 12 (95%CI –0.12 to 0.23×109/L). After adjusting for baseline score in ANCOVA model, in which 32 33 34 13 treatment group and baseline value are independent variables, the least square means estimate for the 35 36 37 14 difference between groups in CRP and blood lymphocyte count is –1.01 mg/L (95%CI –3 to 1 mg/L) 38 39 15 and –0.02×109/L (95%CI –0.16 to 0.13×109/L), respectively. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 22

https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 39

1 2 3 1 DISCUSSION 4 5 2 Principal findings 6 7 8 3 The present study (conducted during the outbreak of COVID–19 in China) is the first randomized, 9 10 11 4 controlled trial showing that there is no increase of negative conversion rate of SARS–CoV–2 12 Confidential: For Review Only 13 5 conferred by the addition of HCQ administration to the current SOC in patients hospitalized mainly 14 15 16 6 with persistent mild to moderate COVID–19. 17 18 7 Comparison with other studies 19 20 21 8 Our negative results on the antiviral efficacy of HCQ obtained in this trial are on the contrary to the 22 23 4,5 24 9 encouraging in–vitro results and the recently reported promising results from a non–randomized 25 26 10 trial with 36 COVID–19 patients.6 It should be noted that participants in our trial had mainly mild to 27 28 29 11 moderate disease with a median 16–day delay of HCQ treatment from symptoms onset. Therefore, 30 31 12 the negative results of our trial are only applicable to patients with persistently mild to moderate 32 33 34 13 COVID–19. The COVID–19 has overwhelmed hospital systems during the pandemic, many of these 35 36 37 14 patients may be treated in the community and may finally end up in hospitalization because of oxygen 38 39 15 needs or even rapid deterioration of the disease. Current data from our trial do not support the use of 40 41 42 16 HCQ in this particular population, particularly considering the increased adverse events (discussed 43 44 17 below). The antiviral efficacy of HCQ remains unknown at the earlier stage, e.g., within 48h of illness 45 46 47 18 onset, the golden window for antiviral treatment in influenza.15 However, it is challenging to conduct 48 49 50 19 such a trial in hospitalized patients and will be easier in outpatient or community settings. The fact 51 52 20 that no restriction of antiviral treatment in our trial should also be considered when interpreting our 53 54 55 21 results. It would be more conclusive to see the antiviral effects of HCQ when comparing a “pure” 56 57 22 control arm. But in the early dangerous outbreak of COVID–19 in China, it is not quite ethical to set 58 59 60 23 a restriction on drugs that could be potentially useful. Nevertheless, the usage of antiviral before and 23

https://mc.manuscriptcentral.com/bmj Page 25 of 39 BMJ

1 2 3 1 after randomization between the HCQ and SOC groups was balanced and therefore might affect little 4 5 2 on our primary endpoint. Moreover, it is not likely to have additional antiviral effects by further 6 7 8 3 escalating the dosage of HCQ used in our trial, because that the dosage of HCQ we choose can reach 9 10 15 11 4 the 50% effective concentrations (EC50) of HCQ against SARS–CoV–2, although we did not 12 Confidential: For Review Only 13 5 monitor the concentration of HCQ in our study. Taken together, future studies could take advantage 14 15 16 6 of our results to design trials in more selective populations, at the earliest stage as possible (<48h of 17 18 7 illness onset), and using more sensitive endpoints, such as viral load shedding. 19 20 21 8 Aside from virus clearance, a numerically greater decline of CRP and an increase of lymphocyte 22 23 24 9 count in HCQ plus SOC group than in the SOC group were observed in our post hoc analysis, 25 26 10 reflecting the anti–inflammatory property HCQ previously established in rheumatic diseases.16 27 28 29 11 However, the absolute changes of CRP or lymphocyte count in both groups were nominal, making 30 31 12 the changes between the groups hard to interpret. From the mechanism perspective, HCQ as a weak 32 33 34 13 base, accumulates within acidic vesicles, such as the lysosomes autophagosomes of phagocytic cells 35 36 16 37 14 and changing local pH value, subsequently inhibits immune activation. Extensive inflammation 38 39 15 response is a hallmark of severe COVID–19.17 However, a reduction in inflammatory mediators 40 41 42 16 might be clinically beneficial (e.g. limiting inflammation) or adverse (e.g. increasing risk of infection) 43 44 17 depending on disease severity and other clinical aspects. Due to the limited number of severe patients, 45 46 47 18 we were not able to investigate the utility of HCQ in relation to disease progression or regression. 48 49 50 19 The similar alleviation rate of clinical symptoms between HCQ and SOC group observed in our trial 51 52 20 suggested limited values of HCQ in controlling local inflammation. However, our current trial was 53 54 55 21 underpowered to reach conclusion on secondary outcomes. These important issues remain to be 56 57 22 investigated in future RCTs. 58 59 60 23 HCQ in our trial was given with a loading dose of 1, 200 mg daily for three days followed by a 24

https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 39

1 2 3 1 maintained dose of 800 mg daily for a total treatment duration of 2 weeks or 3 weeks for 4 5 2 mild/moderate or severe patients, respectively. Serious adverse events occurred in 2 patients and both 6 7 8 3 were reported from HCQ recipients. The overall frequency of adverse events was significantly higher 9 10 11 4 in HCQ recipients than non–recipients. Gastrointestinal events, particularly diarrhea, was most 12 Confidential: For Review Only 13 5 commonly reported, similar to another report using a high dose of HCQ18 Transient blurred vision 14 15 16 6 was reported in one patient whose symptoms recovered 2 day after discontinuation of HCQ. Early 17 18 7 development of retinal damage with a daily dose of 800 to 1,200mg was detected using sensitive 19 20 21 8 retinal screening tests.19 Therefore, the retinal damage could be underestimated in our trial. Events of 22 23 20 24 9 cardiac arrhythmia, e.g., prolonged QT interval was not observed in our trial, possibly due to the 25 26 10 relatively mild/moderate patients investigated or the short–term period of follow–up. However, with 27 28 29 11 the increasing interest of the combined use of HCQ and azithromycin worldwide, physicians should 30 31 12 be cautious of the increased risk of QT interval prolongation and fatal ventricular arrhythmia with 32 33 34 13 azithromycin and other antimicrobials.21,22 Drug–drug interaction10 should be taken into 35 36 37 14 consideration when assessing safety and efficacy endpoints in future HCQ trials. The effects of HCQ 38 39 15 in causing increased levels of digitoxin and metoprolol16 would be particularly relevant in severe 40 41 42 16 COVID–19 patients and therefore would require close monitoring. 43 44 17 Strengths and limitations of this study 45 46 47 18 This study provides the first and timely evidence regarding benefit–risk of HCQ derived from a multi– 48 49 50 19 center randomized controlled trial, which was initiated during the most challenging time of the 51 52 20 COVID–19 outbreak in China. 53 54 55 21 Under such a situation, our study does have several limitations. First, the design of open–label, as 56 57 22 opposed to double–blind design, introduces biased investigator–determined assessments and 58 59 60 23 unbalanced dosage adjustment. Urgent production of placebos mimicking HCQ and the management 25

https://mc.manuscriptcentral.com/bmj Page 27 of 39 BMJ

1 2 3 1 of a multi–center placebo–controlled trial remains challenging during the pandemic. Second, the use 4 5 2 of sequential envelopes is inferior to the interactive web response management system for 6 7 8 3 randomization. Third, the conduction of our trial in the setting of hospitalized patients precludes us 9 10 11 4 from enrolling patients at the early disease stage. In addition, we cannot provide evidence on the 12 Confidential: For Review Only 13 5 utility of HCQ regarding disease progression or regression because our patients are mainly with mild 14 15 16 6 to moderate diseases with only 3 severe patients. Fourth, the results on our main pre–specified 17 18 7 outcomes are not entirely conclusive based on underpowered sample size due to the lack of enough 19 20 21 8 eligible patients to enroll. The recruitment of eligible patients was unexpectedly difficult with almost 22 23 24 9 hundreds of clinical trials launched in the same period in response to the urgent call for the exploration 25 26 10 of effective treatment against COVID–19 by the national health authorities. The rapid decline in 27 28 29 11 eligible new cases owing to the successful containment of COVID–19 in the mid of March 2020 in 30 31 12 China precluded further recruitment to reach our targeted sample size. Fifth, it is difficult to ensure 32 33 34 13 the fidelity to the protocol by site investigators under highly challenging circumstances at the front 35 36 37 14 lines in the COVID–19 treatment centers. Hiring a contract research organization, as we did in our 38 39 15 trial, can greatly support the conduct and oversight of the trial. Sixth, population quarantine of Wuhan 40 41 42 16 and neighboring cities, nationwide travel restrictions, and case/contact isolation were also barriers to 43 44 17 collect and transfer data and paper files. Several prespecified secondary endpoints, including the 45 46 47 18 imaging changes on chest CT, were therefore not finished by the cutoff date of analysis. Finally, the 48 49 50 19 specimens collected in our trial for virus RNA determination are mostly from the upper respiratory 51 52 20 tract rather than bronchoalveolar lavage fluid, which introduced false–negative results.23 But the 53 54 55 21 prespecified definition for virus negative conversion was two negatives with at least 24 hours apart, 56 57 22 which can reduce false negativity. 58 59 60 23 Conclusion and policy implications 26

https://mc.manuscriptcentral.com/bmj BMJ Page 28 of 39

1 2 3 1 The results of our trial did not show additional benefits of virus elimination by adding HCQ to the 4 5 2 current SOC in patients mainly with persistent mild to moderate COVID–19. Adverse events, 6 7 8 3 particularly gastrointestinal events, were more frequently reported in patients receiving HCQ, who 9 10 11 4 were given with a loading dose of 1, 200 mg daily for three days followed by a maintained dose of 12 Confidential: For Review Only 13 5 800 mg daily for remaining days (total treatment duration: 2 weeks or 3 weeks for mild/moderate or 14 15 16 6 severe patients, respectively). Overall, these data do not support adding HCQ to the current SOC in 17 18 7 patients with persistent mild to moderate COVID–19 for eliminating the virus. Our trial may provide 19 20 21 8 initial evidence for the benefit–risk of HCQ and serve as a resource to support further research. 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 27

https://mc.manuscriptcentral.com/bmj Page 29 of 39 BMJ

1 2 3 1 Acknowledgments: 4 5 2 We would like to acknowledge the members of the trial steering committee for their many 6 7 8 3 contributions in conducting the trial under very challenging field conditions, in particular, Prof. 9 10 11 4 Yiping Xu and the members of the independent data and safety monitoring committee (Prof. 12 Confidential: For Review Only 13 5 Shenghan Lai [chair], Prof. Weimin Li, Prof. Hejian Zou, Prof. Xinxin Zhang, and Prof. Gang Chen) 14 15 16 6 for their oversight and the Chief of the Department of Anesthesiology, Ruijin Hospital, Shanghai Jiao 17 18 7 Tong University School of Medicine, Prof. Yan Luo for her support to the trial. We would like to 19 20 21 8 acknowledge the contribution of Xia Chen, Li Li and Jian Li for their support of statistical analysis. 22 23 24 9 We also would like to acknowledge the contribution of the other members of the coordinating centers: 25 26 10 Ming Li, Yongcai Gao, Ming Wu, Zuohan Xiao, Shanshan Deng, Xu Zhang, Xiaobo Chen, Haiyan 27 28 29 11 Hang, Haiguang Xin, Yadong Gao, Dandan Dong, Mingjie, Hou, Qiuping Peng, Yiming Yan, Peng 30 31 12 Gao, Lin Sun, Jingdong Shi, Hong Tang, and Fei Deng. Most importantly, we thank the donation of 32 33 34 13 our investigated drug from the Shanghai Pharmaceuticals Holding Co., Ltd and the patients 35 36 37 14 themselves for their bravery and altruism in participating in this trial. 38 39 15 Funding This work was supported by the Emergent Projects of National Science and Technology 40 41 42 16 (2020YFC0844500), National Natural Science Foundation of China (81970020, 81770025), National 43 44 17 Key Research and Development Program of China (2016YFC0901104), Shanghai Municipal Key 45 46 47 18 Clinical Specialty (shslczdzk02202, shslczdzk01103), National Innovative Research Team of High– 48 49 50 19 level Local Universities in Shanghai, Shanghai Key Discipline for Respiratory Diseases 51 52 20 (2017ZZ02014), National Major Scientific and Technological Special Project for Significant New 53 54 55 21 Drugs Development (2017ZX09304007), Key Projects in the National Science and Technology Pillar 56 57 22 Program during the Thirteenth Five–year Plan Period (2018ZX09206005–004, 2017ZX10202202– 58 59 60 23 005–004, 2017ZX10203201–008). The funders played no role in study design, data collection, data 28

https://mc.manuscriptcentral.com/bmj BMJ Page 30 of 39

1 2 3 1 analysis, data interpretation, or reporting. The guarantors had full access to all the data in the study, 4 5 2 take responsibility for the integrity of the data and the accuracy of the data analysis, and had final 6 7 8 3 responsibility for the decision to submit for publication. 9 10 11 4 Competing interests: All authors have completed the ICMJE uniform disclosure form at 12 Confidential: For Review Only 13 5 www.icmje.org/coi_disclosure.pdf and declare: no support from any organization for the submitted 14 15 16 6 work; no financial relationships with any organizations that might have an interest in the submitted 17 18 7 work in the previous three years; no other relationships or activities that could appear to have 19 20 21 8 influenced the submitted work. 22 23 24 9 Ethical approval: The trial was approved by the Institutional Review Board of Ruijin Hospital, 25 26 10 Shanghai Jiao Tong University School of Medicine (KY2020–29). 27 28 29 11 Patient consent: Written informed consent was obtained. 30 31 12 Data sharing: Anonymized datasets can be made available on reasonable request after approval from 32 33 34 13 the trial management committee and after signing a data access agreement. Proposals should be 35 36 37 14 directed to the corresponding author. 38 39 15 Transparency declaration: The lead author (the manuscript’s guarantor) affirms that the manuscript 40 41 42 16 is an honest, accurate, and transparent account of the study being reported; that no important aspects 43 44 17 of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, 45 46 47 18 registered) have been explained. 48 49 50 19 Dissemination to participants and related patient and public communities: No patients were 51 52 20 involved in setting the research question or the outcome measures, nor were they involved in 53 54 55 21 developing plans for recruitment, design or implementation of the study. No patients were asked to 56 57 22 advise on interpretation or writing up of results. There are no plans to disseminate the results of the 58 59 60 23 research to study participants or the relevant patient community. 29

https://mc.manuscriptcentral.com/bmj Page 31 of 39 BMJ

1 2 3 1 Reference 4 5 2 1. Coronavirus COVID-19 Global Cases by Johns Hopkins CSSE. 6 3 https://www.arcgis.com/apps/opsdashboard/index.html#/85320e2ea5424dfaaa75ae62e5c06e61 7 4 (accessed April 22 2020). 8 9 5 2. Dyall J, Gross R, Kindrachuk J, et al. Middle East Respiratory Syndrome and Severe Acute 10 6 Respiratory Syndrome: Current Therapeutic Options and Potential Targets for Novel Therapies. 11 7 Drugs 2017; 77: 1935-66. 12 Confidential: For Review Only 8 3. Cao B, Wang Y, Wen D, et al. A Trial of Lopinavir-Ritonavir in Adults Hospitalized with Severe 13 14 9 Covid-19. N Engl J Med 2020; published online March 18. https://doi.org/10.1056/NEJMoa2001282. 15 10 4. Wang M, Cao R, Zhang L, et al. Remdesivir and chloroquine effectively inhibit the recently 16 11 emerged novel coronavirus (2019-nCoV) in vitro. Cell Res 2020; 30: 269-71. 17 18 12 5. Yao X, Ye F, Zhang M, et al. In Vitro Antiviral Activity and Projection of Optimized Dosing 19 13 Design of Hydroxychloroquine for the Treatment of Severe Acute Respiratory Syndrome 20 14 Coronavirus 2 (SARS-CoV-2). Clin Infect Dis 2020; published online Mar 9. 21 22 15 https://doi.org/10.1093/cid/ciaa237. 23 16 6. Gao J, Tian Z, Yang X. Breakthrough: Chloroquine phosphate has shown apparent efficacy in 24 17 treatment of COVID-19 associated pneumonia in clinical studies. Biosci Trends 2020; 14: 72-3. 25 18 7. National Health Commission of the People’s Republic of China. Chinese guideline for the 26 27 19 management COVID-19 (version 7.0, in Chinese). 2020. 28 20 http://www.nhc.gov.cn/yzygj/s7653p/202003/46c9294a7dfe4cef80dc7f5912eb1989/files/ce3e69458 29 21 32a438eaae415350a8ce964.pdf. (accessed March 30 2020). 30 31 22 8. EUA Hydroxychloroquine sulfate Health Care Provider Fact Sheet. 32 23 https://www.fda.gov/emergency-preparedness-and-response/mcm-legal-regulatory-and-policy- 33 24 framework/emergency-use-authorization (accessed March 31 2020). 34 35 25 9. 'This is insane!' Many scientists lament Trump's embrace of risky malaria drugs for coronavirus. 36 26 https://www.sciencemag.org/news/2020/03/insane-many-scientists-lament-trump-s-embrace-risky- 37 27 malaria-drugs-coronavirus# (accessed March 30 2020). 38 28 10. Rainsford KD, Parke AL, Clifford-Rashotte M, Kean WF. Therapy and pharmacological 39 40 29 properties of hydroxychloroquine and chloroquine in treatment of systemic lupus erythematosus, 41 30 rheumatoid arthritis and related diseases. Inflammopharmacology 2015; 23: 231-69. 42 31 11. Baden LR, Rubin EJ. Covid-19 - The Search for Effective Therapy. The New England journal of 43 44 32 medicine 2020; published online. https://doi.org/10.1056/NEJMe2005477: 1-2. 45 33 12. National Health Commission of the People’s Republic of China. Chinese guideline for the 46 34 management COVID-19 (version 5.0, in Chinese). 2020. 47 48 35 http://www.nhc.gov.cn/yzygj/s7653p/202002/d4b895337e19445f8d728fcaf1e3e13a/files/ab6bec7f9 49 36 3e64e7f998d802991203cd6.pdf. (accessed February 9 2020). 50 37 13. Chen T, Wu D, Chen H, et al. Clinical characteristics of 113 deceased patients with coronavirus 51 38 disease 2019: retrospective study. BMJ 2020; published online Mar 26. 52 53 39 https://doi.org/10.1136/bmj.m1091. 54 40 14. Vickers AJ, Altman DG. Statistics notes: Analysing controlled trials with baseline and follow up 55 41 measurements. BMJ 2001; 323: 1123-4. 56 57 42 15. Harper Scott A, Bradley John S, Englund Janet A, et al. Seasonal Influenza in Adults and 58 43 Children—Diagnosis, Treatment, Chemoprophylaxis, and Institutional Outbreak Management: 59 44 Clinical Practice Guidelines of the Infectious Diseases Society of America. Clin Infect Dis 2009; 48: 60 30

https://mc.manuscriptcentral.com/bmj BMJ Page 32 of 39

1 2 1 1003-32. 3 2 16. Schrezenmeier E, Dorner T. Mechanisms of action of hydroxychloroquine and chloroquine: 4 5 3 implications for rheumatology. Nat Rev Rheumatol 2020; 16: 155-66. 6 4 17. Guan W, Ni Z, Hu Y, et al. Clinical Characteristics of Coronavirus Disease 2019 in China. N 7 5 Engl J Med 2020; published online Feb 28. https://doi.org/10.1056/NEJMoa2002032. 8 9 6 18. Munster T, Gibbs JP, Shen D, et al. Hydroxychloroquine concentration-response relationships in 10 7 patients with rheumatoid arthritis. Arthritis Rheum 2002; 46: 1460-9. 11 8 19. Leung L, Neal J, Wakelee H, Sequist L, Marmor M. Rapid Onset of Retinal Toxicity From High- 12 Confidential: For Review Only 9 Dose Hydroxychloroquine Given for Cancer Therapy. Am J Ophthalmol 2015; 160: 799-805.e1. 13 14 10 20. van Roon EN, van den Bemt PM, Jansen TL, Houtman NM, van de Laar MA, Brouwers JR. An 15 11 evidence-based assessment of the clinical significance of drug-drug interactions between disease- 16 12 modifying antirheumatic drugs and non-antirheumatic drugs according to rheumatologists and 17 18 13 pharmacists. Clin Ther 2009; 31: 1737-46. 19 14 21. Ray W, Murray K, Meredith S, Narasimhulu S, Hall K, Stein C. Oral erythromycin and the risk 20 15 of sudden death from cardiac causes. N Engl J Med 2004; 351: 1089-96. 21 22 16 22. Owens RC, Jr., Nolin TD. Antimicrobial-associated QT interval prolongation: pointes of interest. 23 17 Clin Infect Dis 2006; 43: 1603-11. 24 18 23. Wang W, Xu Y, Gao R, et al. Detection of SARS-CoV-2 in Different Types of Clinical 25 19 Specimens. Jama 2020; published online Mar 11. https://doi.org/10.1001/jama.2020.3786. 26 27 20 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 31

https://mc.manuscriptcentral.com/bmj Page 33 of 39 BMJ

1 2 3 1 Figure legend 4 5 6 2 Figure 1. Screening, Randomization, and Follow–up. Shown is the disposition of the trial participants. 7 8 9 10 3 Figure 2. Kaplan–Meier curves of the time to negative conversion of SARS–CoV–2 on RT–PCR test 11 12 4 in SOC plusConfidential: HCQ versus SOC in the intention–to–treat For Review population. Shown Only are data for 75 patients 13 14 15 5 assigned to SOC plus HCQ and 75 assigned to SOC. The overall negative conversion rate was 85.4% 16 17 18 6 (95% CI 73.8% to 93.8%), similar to that of the SOC group 81.3% (95%CI 71.2% to 89.6%) within 19 20 7 28–day (p=0.34). Between-group difference was 4.1% (95%CI –10.3% to 18.5%). The median time 21 22 23 8 to negative conversion was also similar in the SOC plus HCQ group (8 days, 95%CI 5 to 10 days) 24 25 26 9 with that in the SOC group (7 days, 95%CI 5 to 8 days). Data from patients who did not have negative 27 28 29 10 conversion were censored (tick marks) at the last visit date. RT–PCR=real time reverse transcription 30 31 11 polymerase chain reaction; SOC=standard–of–care; HCQ=hydroxychloroquine; CI=confidence 32 33 34 12 interval. 35 36 37 38 13 Figure 3 Kaplan–Meier curves of the time to alleviation of clinical symptoms in SOC plus HCQ 39 40 14 versus SOC in the intention–to–treat population. Shown are data for 55 symptomatic patients assigned 41 42 43 15 to SOC plus HCQ and 64 symptomatic patients assigned to SOC. The overall rate of symptoms 44 45 46 16 alleviation within 28–day was similar (p=0.97) between patients with SOC with (59.9%, 95%CI 45.0% 47 48 49 17 to 75.3%) and without HCQ (66.6%, 95%CI 39.5% to 90.9%). Between-group difference was –6.6% 50 51 18 (95%CI –41.3% to 28.0%). The median time to alleviation of clinical symptoms was similar in the 52 53 54 19 SOC plus HCQ group with that in the SOC group (19 days versus 21 days). Data from patients who 55 56 57 20 did not have symptoms alleviation were censored (tick marks) at the last visit date. SOC=standard– 58 59 21 of–care; HCQ=hydroxychloroquine; CI=confidence interval. 60 32

https://mc.manuscriptcentral.com/bmj BMJ Page 34 of 39

1 2 3 1 Table 1. Baseline Demographic and Clinical Characteristics of the Patients in the Intention–to–Treat Population. 4 5 SOC plus HCQ SOC Total Characteristics* 6 (N=75) (N=75) (N=150) 7 8 Age — yr 48.0±14.1 44.1±15.0 46.1±14.7 9 Male sex — no. (%) 42 (56.0) 40 (53.3) 82 (54.7) 10 Body mass index (% with missing data)† 23.9±3.24 (1.3) 23.2±3.0 (5.3) 23.5±3.2 (3.3) 11 Days from disease onset to randomization (% 12 Confidential: For16.0±9.9 Review (2.7) 17.1±11.1 Only (1.3) 16.6±10.5 (2.0) 13 with missing data) 14 Exposure history — no./total no. (%) 15 16 Hubei province exposure 50/72 (69.4) 53/71 (74.6) 103/143 (72) 17 Contact with confirmed COVID–19 patient 18 39/72 (54.2) 32/71 (45.1) 71/143 (49.7) (s) 19 20 Others 1/72 (1.4) 1/71 (1.4) 2/143 (1.4) 21 No exposure 2/72 (2.8) 9/71 (12.7) 11/143 (7.7) 22 Unknown 5/72 (6.9) 5/71 (7) 10/143 (7) 23 24 Medication prior to randomization — no. (%) 47 (62.7) 43 (57.3) 90 (60.0) 25 Antivirals 28 (37.3) 24 (32.0) 52 (34.7) 26 27 Arbidol 12 (16.0) 8 (10.7) 20 (13.3) 28 Lopinavir–ritonavir 18 (24.0) 14 (18.7) 32 (21.3) 29 Oseltamivir 3 (4.0) 3 (4.0) 6 (4.0) 30 31 Entecavir 1 (1.3) 0 1 (0.7) 32 Virazole 3 (4.0) 6 (8.0) 9 (6.0) 33 Ganciclovir 0 2 (2.7) 2 (1.3) 34 35 Disease severity — no. (%) 36 Mild 15 (20.0) 7 (9.3) 22 (14.7) 37 Moderate 59 (78.7) 67 (89.3) 126 (84.0) 38 39 Severe 1 (1.3) 1 (1.3) 2 (1.3) 40 Coexisting conditions — no./total no. (%) 28 (37.3) 17 (22.7) 45 (30.0) 41 42 Diabetes 12 (16.0) 9 (12.0) 21 (14.0) 43 Hypertension 6 (8.0) 3 (4.0) 9 (6.0) 44 Others 21 (28.0) 10 (13.3) 31 (20.7) 45 46 Vital Signs (%with missing data) 47 Body temperature — °C 36.9±0.47 (4) 36.8±0.48 (0.0) 36.8±0.5 (2.0) 48 Pulse — beats/min 82.75±8.0 (2.7) 82.5±9.4 (5.3) 82.6±8.7 (4.0) 49 50 Respiratory rate — breaths/min 19.6±1.3 (2.7) 19.7±1.7 (6.7) 19.6±1.5 (4.7) 51 Systolic blood pressure — mm Hg 126.3±13.2 (6.7) 123.5±11.2 (8.0) 124.9±12.3 (7.3) 52 Diastolic blood pressure — mm Hg 79.1±8.5 (6.7) 76.8±8.0 (8.0) 77.9±8.3 (7.3) 53 54 Pulse oximetry — % 97.4±1.6 (0) 97.3±1.6 (2.7) 97.4±1.6 (1.3) 55 Symptoms — no./total no. (%) 56 Fever 43/72 (59.7) 40/75 (53.3) 83/157 (52.9) 57 58 Cough 35/68 (51.5) 26/68 (38.2) 61/136 (44.9) 59 Sputum production 11/68 (16.2) 4/68 (5.9) 15/136 (11) 60 33

https://mc.manuscriptcentral.com/bmj Page 35 of 39 BMJ

1 2 Shortness of breath 15/68 (22.1) 4/68 (5.9) 19/136 (14) 3 Nasal congestion 0 (0) 0 (0) 0 (0) 4 5 Pharynx discomfort 2/68 (2.9) 4/68 (5.9) 6/136 (4.4) 6 Fatigue 5/68 (7.4) 1/68 (1.5) 6/136 (4.4) 7 Laboratory parameters (% with missing data) 8 9 White cell count — ×109/liter 5.59±1.9 (0) 5.6±1.8 (0.0) 5.6±1.8 (0.0) 10 Lymphocyte count — ×109/liter 1.46±0.6 (0) 1.6±0.5 (0.0) 1.5±0.57 (0.0) 11 9 12 NeutrophilConfidential: count— ×10 /liter For3.55±1.6 Review (0) 4.2±6.2 Only (0.0) 3.9±4.51 (0.0) 13 Platelet count — ×109/liter 214.8±68.1 (0) 211.7±71.6 (0.0) 213.2±69.7 (0.0) 14 Hemoglobin —g/liter 128.8±17.5 (0) 129.1±17.1 (0.0) 129.0±17.3 (0.0) 15 16 Aspartate aminotransferase — U/liter 25.0±13.5 (0) 26±14.7 (0.0) 25.5±14.1 (0.0) 17 Alanine aminotransferase — U/liter 31.4±26.3 (0) 32.7±25.2 (1.3) 32.1±25.7 (0.7) 18 γ–glutamyl transpeptidase — U/liter 46.9±61.8 (2.7) 44.0±51.8 (2.7) 45.4±56.9 (2.7) 19 20 Total bilirubin — μmol/liter 11.6±8.4 (1.3) 12.8±7.7 (2.7) 12.2±8.1 (2.0) 21 Albumin — g/L 39.9±4.5 (1.3) 40.4±4.4 (1.3) 40.1±4.4 (1.3) 22 Lactate dehydrogenase — U/liter 203.9±65.2 (12) 190.9±49.5 (10.7) 197.4±58.0 (11.3) 23 24 Creatine kinase — U/liter 74.4±110.1 (10.7) 71.0±52.6 (9.3) 72.7±85.7 (10.0) 25 Creatine kinase isoenzyme–MB — U/liter 8.0±4.2 (38.7) 6.8±3.9 (41.3) 7.4±4.0 (40.0) 26 Creatinine — μmol/liter 71.2±38.4 (1.3) 63.9±16.0 (1.3) 67.5±29.5 (1.3) 27 28 Blood urea nitrogen — mmol/liter 3.5±1.0 (41.3) 3.1±0.7 (48.0) 3.3±0.9 (44.7) 29 Urea — mmol/liter 4.0±3.0 (58.7) 3.8±1.2 (57.3) 4.0±2.2 (58.0) 30 31 International normalized ratio 1.0±0.1 (2.7) 1.0±0.1 (1.3) 1.0±0.1 (2.0) 32 C–reactive protein mg/liter 9.9±13.3 (2.7) 7.4±12.8 (1.3) 8.6±13.1 (2.0) 33 Erythrocyte sedimentation rate 30.6±28.6 (4) 25.4±21.7 (5.3) 28.0±25.4 (4.7) 34 35 TNF–α — pg/milliliter 4.9±4.1 (90.7) 4.8±3.6 (90.7) 4.8±3.7 (90.7) 36 IL–6 — pg/milliliter 12.9±36.3 (58.7) 8.9±13.0 (61.3) 11.0±27.4 (60.0) 37 1 * Plus–minus values are means ±SD. SOC=standard–of–care, HCQ= hydroxychloroquine, COVID–19=coronavirus 38 39 2 disease 2019. To convert the values for creatinine to milligrams per deciliter, divide by 88.4. To convert the values 40 3 for total bilirubin to milligrams per deciliter, divide by 17.1. 41 † The body mass index is the weight in kilograms divided by the square of the height in meters. 42 4 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 34

https://mc.manuscriptcentral.com/bmj BMJ Page 36 of 39

1 2 3 1 Table 2. Treatments after Randomization in Patients in the Intention–to–Treat Population. 4 5 SOC plus HCQ SOC Total 6 7 Treatments 8 (N=75) (N=75) (N=150) 9 10 11 Antiviral — no. (%) 47 (62.7) 48 (64.0) 95 (63.3) 12 Confidential: For Review Only 13 Arbidol 37 (49.3) 33 (44.0) 70 (46.7) 14 15 16 Virazole 13 (17.3) 15 (20.0) 28 (18.7) 17 18 Lopinavir–Ritonavir 13 (17.3) 12 (16.0) 25 (16.7) 19 20 21 Oseltamivir 8 (10.7) 9 (12.0) 17 (11.3) 22 23 24 Entecavir 1 (1.3) 1 (1.3) 2 (1.3) 25 26 Antibiotics — no. (%) 32 (42.7) 27 (36.0) 59 (39.3) 27 28 29 Systemic glucocorticoid therapy — no. (%) 6 (8.0) 4 (5.3) 10 (6.7) 30 31 2 * SOC=standard–of–care, HCQ= hydroxychloroquine 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 35

https://mc.manuscriptcentral.com/bmj Page 37 of 39 BMJ

1 2 3 1 Table 3. Summary of Adverse Events in the Safety Population. 4 SOC plus HCQ SOC 5 Adverse Events* P-value 6 (N=70) (N=80) 7 no. of patients (%) 8 9 Any adverse event 21 (30) 7 (8.8) 0.001 10 Serious adverse event 2 (2.9) 0 0.216 11 Disease progression 1 (1.4) 0 0.467 12 Confidential: For Review Only Upper respiratory tract infection 1 (1.4) 0 0.467 13 14 Non serious adverse event 19 (27.1) 7 (8.8) 0.004 15 Diarrhea 7 (10.0) 0 0.004 16 Vomiting 2 (2.9) 0 0.216 17 18 Nausea 1 (1.4) 0 0.467 19 Abdominal discomfort 1 (1.4) 0 0.467 20 Thirst 1 (1.4) 0 0.467 21 22 Abdominal bloating 0 1 (1.3) 1 23 Sinus bradycardia 1 (1.4) 0 0.467 24 Hypertension 1 (1.4) 0 0.467 25 Orthostatic hypotension 1 (1.4) 0 0.467 26 27 Hypertriglyceridemia 1 (1.4) 0 0.467 28 Decreased appetite 1 (1.4) 0 0.467 29 Fatigue 1 (1.4) 0 0.467 30 31 Fever 0 1 (1.3) 1 32 Dyspnea 1 (1.4) 0 0.467 33 Flush 1 (1.4) 0 0.467 34 35 Liver abnormality 0 1 (1.3) 1 36 Kidney injury 1 (1.4) 0 0.467 37 Coagulation dysfunction 1 (1.4) 0 0.467 38 Hepatic steatosis 0 1 (1.3) 1 39 40 Otitis externa 0 1 (1.3) 1 41 Blurred vision 1 (1.4) 0 0.467 42 Decreased white blood cell 1 (1.4) 0 0.467 43 44 Increased alanine aminotransferase 1 (1.4) 1 (1.3) 1 45 Increased serum amylase 1 (1.4) 0 0.467 46 Decreased neutrophil count 1 (1.4) 0 0.467 47 48 Increased serum amyloid A 0 1 (1.3) 1 49 2 * Multiple occurrences of the same adverse event in one patient were counted. SOC=standard–of–care, 50 3 HCQ=hydroxychloroquine. P-value was calculated using Chi–square test followed by Fisher’s exact test as appropriate. 51 52 53 4 54 55 56 57 5 58 59 60 36

https://mc.manuscriptcentral.com/bmj BMJ Page 38 of 39

191 Participants were assessed 1 2 for eligibility 3 4 41 Did not meet eligibility 5 Confidential: For Review Only 6 criteria 7 8 9 150 Underwent randomization 10 11 12 13 14 75 Were assigned to the 75 Were assigned to the 15 SOC plus HCQ group SOC group 16 17 18 19 20 21 Intention-to-treat population Intention-to-treat population 22 23 24 25 6 Did not receive 26 HCQ 27 28 29 1 Received HCQ 30 31 32 33 34 70 Were included in 80 Were included in 35 the safety population the safety population 36 37 38 39 https://mc.manuscriptcentral.com/bmj 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 PageConfidential: 39 of 39 ForBMJ Review Only 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Patients w ith positive SAR S-C oV-2 on R T-PC R test (% ) p=0.34 by log-rank 19 20 21 22 https://mc.manuscriptcentral.com/bmj 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Page 40 of 39

Confidential: For Review Onlylog-rank by p=0.97 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Cumulative rate of symptoms alleviation (%) alleviation symptoms of rate 16 Cumulative 17 18 19 20 21 22 https://mc.manuscriptcentral.com/bmj 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60