Abstract
Importance No clinically proven effective antiviral strategy exist for the emerging epidemic Coronavirus Disease 2019 (COVID-19).
Objective To compare the efficacy and safety of favipiravir and arbidol to treat COVID-19 patients.
Design Prospective, multicenter, open-label, randomized superiority trial in February, 2020.
Setting Multicenter study.
Participants Patients with confirmed COVID-19 admitted to 3 hospitals from Feb. 20, 2020 to Mar. 12, 2020.
Interventions Conventional therapy + favipiravir or arbidol.
Main Outcomes and Measures The primary outcome was clinical recovery rate of day 7. Duration of fever, cough relief latency, and auxiliary oxygen therapy or noninvasive mechanical ventilation rate were the secondary outcomes. The patients with chest CT imaging and laboratory-confirmed COVID-19 infection, aged 18 years or older were randomly assigned to receive favipiravir or arbidol. Safety data were collected for further follow-up for a week.
Results 120 patients were assigned to favipiravir group (116 assessed) and 120 to arbidol group (120 assessed). In full analysis set (FAS) cohort, clinical recovery rate of day 7 does not significantly differ between the favipiravir group (61.20% (71/116)) and the arbidol group (51.67% (62/120)) (P=0.1396, OR: 1.47). The latency to fever reduction and cough relief in favipiravir group was significantly shorter than that in arbidol group (both P<0.0001). No statistical difference was observed of auxiliary oxygen therapy or noninvasive mechanical ventilation rate (both P>0.05). The most frequently observed favipiravir-associated adverse events were abnormal LFT (10/116, 8.62%, OR: 0.86, P=0.7156), psychiatric symptom reactions (5/116, 4.31%, OR: 5.17, P=0.1149), digestive tract reactions (16/116, 13.79%, OR: 0.97, P=0.6239) and raised serum uric acid (16/116, 13.79%, OR: 5.52, P=0.0014).
Conclusions and Relevance Among patients with confirmed COVID-19, favipiravir, compared to arbidol, did not significantly improve the clinically recovery rate by 7 days. Favipiravir significantly improved time-to-relief for fever and cough. Antiviral-associated adverse effects associated with favipiravir are mild and manageable.
Trial Registration This study is registered with Chictr.org.cn, number ChiCTR2000030254.
Question How about the efficacy and safety of favipiravir to treat COVID-19 patients?
Findings Among patients with confirmed COVID-19, favipiravir, compared to arbidol, did not significantly improve the clinically recovery rate by 7 days. Favipiravir treatment significantly improved time-to-relief for fever and cough, and is associated with manageable adverse effects.
Meaning For COVID-19 patients, favipiravir is a safe and at least similarly effective option compared to arbidol.
Introduction
In Dec. 2019, an outbreak of pneumonia caused by SARS-CoV-2 occurred in Wuhan, followed has spread rapidly throughout China. As of Mar. 14, 2020, the WHO reported 146,181 confirmed cases across more than 130 countries.1 The global mortality rate of COVID-19 is 3.4%,2 whereas its mortality rate in Wuhan is 4.3%,3 as the proportion of severe/critical cases in Wuhan is relatively high.
At the moment, there is no specific treatment for COVID-19, so healthcare providers treated the clinical symptoms (fever and breathing difficulties) of patients. Supportive care such as fluid management and anxiliary oxygen therapy can be highly effective for patients. Specifically, no antiviral drugs with RCT proven clinical efficacy for COVID-19 were reported. Clinical studies of some drugs (remdesivir, human interferon alpha-2b, ribavirin, chloroquine, lopinavir and arbidol) were undergoing to test the efficacy and safety for treatment of COVID-19.4
SARS-CoV-2 and influenza viruses have a similar disease presentation. Both clinical manifestations are dominated by respiratory symptoms, presenting a wide range of illness from asymptomatic, mild to severe/critical disease with high mortality rate. Importantly, coronavirus and influenza viruses are both RNA virus which depend on RNA-dependent RNA polymerase (RdRp) to replicate. Although no antiviral drugs were proven effective for COVID-19, clinically assessible antiviral options exist for influenza. Arbidol is an antiviral treatment for influenza infection used in Russia and China and has been proposed as a potential treatment for COVID-19.5,6 Although arbidol is currently widely used in China for treating COVID-19 based on preliminary evidences, no RCT is available to show its efficacy for COVID-19. Also, favipiravir, an antiviral drug targeting RdRP,7 approved in Japan for influenza, has an IC50 of 0.013-0.48 ug/ml for influenza A, comparing with 2.7-13.8 ug/ml of arbidol.8 Hence, we consider favipiravir might serve as a potential candidate to treat COVID-19.
We hypothesized that favipiravir would be superior to arbidol in terms of improving clinical recovery rate, and alleviating fever, cough, and breathing difficulties compared with arbidol. Therefore, we assessed the clinical efficacy and safety of favipiravir versus arbidol as treatment for COVID-19.
Methods
Study design and participants
We conducted a prospective, multicenter, open-labelled, randomized superiority trial in 240 patients with COVID-19 pneumonia at three hospitals (120 from Zhongnan Hospital of Wuhan University, 88 from Wuhan Leishenshan Hospital, 32 from The Third People’s Hospital of Hubei Province). Patients were prospectively enrolled and followed-up from Feb. 20 to Mar. 12, 2020 (Figure 1). According to the proportion of 1:1 between the experimental group (favipiravir) and the control group (arbidol), the randomized open label was produced by professional statistical software SAS9.4. Additional details are in the Trial Protocol (Supplementary file 1), approved by the Institutional Ethics Committee (Nr. 2020040). Written informed consents were obtained from all participants or their authorized representatives.
According to the Chinese Diagnosis and Treatment Protocol for Novel Coronavirus Pneumonia at that moment,9,10 COVID-19 could be diagnosed without a positive nucleic acid test result by: (1) a positive chest CT scan; (2) significant clinical manifestation including fever, cough, breath difficulty and other indications of viral infection of lower respiratory tract; and (3) laboratory results indicating lymphopenia and (optional) leukopenia.
Eligibility criteria was defined as: (1) Aged 18 years or older; (2) Voluntarily signed informed consent; (3) Initial symptoms were within 12 days; (4) Diagnosed as COVID-19 pneumonia.
Patients were excluded if they meet any of following criteria: (1) Allergic to favipiravir or arbidol; (2) ALT/AST increased to over 6 times of normal upper range, or with Child-Pugh C; (3) Severe/critical patients whose expected survival time <48 hours; (4) Female in pregnancy; (5) HIV infection; (6) Considered unsuitable by researchers.
Procedures
Arbidol is the recommended drug in Chinese guideline.10 The experimental group (favipiravir) was treated with routine treatment+favipiravir tablets (1600mg each time on the first day, twice a day; 600mg each time from the second day to the end of the experiment, twice a day). The control group (arbidol) was treated with routine therapy+arbidol (200mg each time, 3 times a day, from the first day to the end of the trial). The course of treatment in both groups was 7-10 days. If necessary, the treatment time could be extended to 10 days according to the judgment of researchers.
Definitions
The primary outcome was the clinical recovery rate at 7 days or the end of treatment. Clinical recovery was defined as continuous (>72 hours) recovery of body temperature, respiratory rate, oxygen saturation and cough relief after treatment, with following quantitative criteria: axillary temperature ≤36.6°C; respiratory frequency ≤24 times/min; Oxygen saturation ≥98% without oxygen inhalation; mild or no cough. The armpit temperature, respiratory rate, oxygen saturation without oxygen, oxygen therapy and noninvasive positive pressure ventilation (NPPV) were recorded in daily follow-up. Repeated measurements were made at least twice in each follow-up. The measurements were taken after 15 minutes rest at room temperature (23±2°C).
Secondary outcomes included the time from randomization to fever reduction (patients with fever at the time of enrollment), the time from randomization to cough relief (patients with moderate or severe cough at the time of enrollment), the rate of auxiliary oxygen therapy or noninvasive mechanical ventilation during the trial, the all-cause mortality during the trial, the rate of respiratory failure during the trial (defined as SPO2 ≤90% without oxygen inhalation or PaO2/FiO2 <300mmHg, requires oxygen therapy or higher respiratory support).
Blood biochemistry, urinalysis, coagulation function, C-reactive protein and nucleic acid were examined on the third day (D3±1 day). The above index and CT were examined on the seventh day (D7±2 day) after taking the drug, and the adverse events and concomitant medication were observed. Classification criteria of moderate, severe and critical COVID-19 patients was described in Trial Protocol.
Statistical Analysis
Sample size estimation: the expected clinical recovery rate at day 7 of the experimental group is 70%, the clinical recovery of the control group is 50%, α=0.025 (single side), β=0.20, power=0.80. According to the distribution ratio of 1:1 between the experimental group and the control group, the statistical sample size is 92 participants in each group. The sample size increased about 20% considering factors such as shedding/elimination. The trial was designed to include 240 participants in the group, including 120 in the experimental group and 120 in the control group.
SAS9.4 software was used for statistical analysis. For the main efficacy indicator/primary outcome (clinical recovery rate after 7 days or the end of treatment), the comparison between the experimental group and the control group adopts the optimal test. We calculated the bilateral 95% CI of the difference between the clinical recovery rate of the experimental group and the control group. If the lower limit was larger than 0, it was considered the experimental group is superior to the control group. Log rank test was used to compare the recovery latency between the two groups. For the secondary efficacy indicators/secondary outcomes, student’s t-test or Wilcoxon rank sum test (if t-test was not applicable) was performed for safety indicators and continuous variables, Wilcoxon rank sum test was used for grade variables. Frequency or composition (%) were used for statistical description of categorial variables, and Chi-square test or Fisher’s exact test was used for comparison between groups. For all statistical tests, P<0.05 (bilateral) were considered as statistically significant.
Results
Basic characteristics of patients
A total of 236 patients with COVID-19 were enrolled in FAS (Figure S1-S2), 116 in the favipiravir group and 120 in the arbidol group (Table 1). In the favipiravir group, 59 (50.86%) were males and 57 (49.14%) were females, 87 (75.00%) were <65 years and 29 (25.00%) were ≥65 years, 36 (31.03%) were with hypertension and 14 (12.07%) with diabetes. In the arbidol group, 51 (42.50%) were males and 69 (57.50%) were females, 79 (65.83%) were <65 years and 41 (34.17%) were ≥65 years, 30 (25.00%) were with hypertension and 13 (10.83%) with diabetes.
Basic characteristics of the participants.
Main signs and symptoms for patients enrolled were fever, fatigue, dry cough, myalgia, dyspnea, expectoration, sore throat, diarrhea, dizziness, insomnia and conjunctivitis, none of which were significantly different between groups. Neither that nucleic acid tests for COVID-19, lymphocyte count, erythrocyte sedimentation rate and C-reactive protein differed between groups. 116 cases in favipiravir group and 119 in arbidol group underwent chest CT, 112 (96.55%) and 114 (95.80%) were diagnosed COVID-19 pneumonia according to the diagnostic criteria (P=0.7635). Overall, no significant difference of basic characteristics of patients between the two groups was observed. However, we noticed a marginally increased ratio of severe to critical patients in the favipiravir group (16+2) compared to arbidol group (8+1) (P=0.0658, Fisher’s exact test, OR: 2.25 [0.91-5.98]).
Comparison of clinical recovery rate of day 7 of favipiravir and arbidol in COVID-19 patients
The group statistics of primary and secondary outcomes were presented in Table 2 and Table S1. At day 7, 62/120 (51.67%) in the arbidol group and 71/116 (61.21%) patients in the favipiravir group (P=0.1396) were clinically recovered, with the difference of recovery rate between two groups (95% CI) was 0.0954 (−0.0305, 0.2213). Hence, we conclude that favipiravir does not show superior efficacy compared to arbidol in terms of improving the clinical recovery rate at day 7.
Comparison of clinical recovery rate of day 7.
Post-hoc test for interaction between treatment and clinical classification showed no interaction between these two factors, both of which contributed to the primary outcome (P=0.017 for treatment, and P<0.001 for clinical classification, with a general linear model). A post-hoc analysis found that for moderate patients with COVID-19, clinical recovery of day 7 was 62/111 (55.86%) in the arbidol group and 70/98 (71.43%) in the favipiravir group (P=0.0199), with the difference of recovery rate between two groups (95% CI) was 0.1557 (0.0271, 0.2843); for severe/critical patients, clinical recovery rate was 0/9 (0%) in the arbidol group and 1/18 (5.56%) in the favipiravir group (P=0.4712), with the difference of recovery rate between two groups (95% CI) was 0.0556 (−0.0503, 0.1614).
Comparison of duration of fever, cough relief time and auxiliary oxygen therapy or noninvasive mechanical ventilation rate
Table 3 displayed duration of fever, cough relief time and auxiliary oxygen therapy or noninvasive mechanical ventilation rate between the favipiravir and arbidol groups. During the course of trial observation, for patients in favipiravir group, 71/116 (61.2%) had a fever and 78/116 (67.2%) had a cough; for patients in the arbidol group, 74/120 (61.7%) had a fever and 73/120 (60.8%) had a cough. Whilst the incidence of fever and cough did not differ between two groups at baseline, both the latency to fever reduction and cough relief in the favipiravir group was significantly shorter than that in the arbidol group (P<0.0001).
Comparison of duration of fever, cough relief time and other secondary outcomes.
The incidence of de novo auxiliary oxygen therapy (AOT) or noninvasive mechanical ventilation (NMV) was 27/120 (22.50%) in the arbidol group and 21/116 (18.10%) in the favipiravir group (P=0.4015), with the difference of incidence rate between 2 groups (95% CI) was −4.40% (−14.64%, 5.85%). For all cases enrolled in this study, the all-cause mortality was 0. The number of cases of respiratory failure were 4 in arbidol group and 1 in favipiravir group (P=0.3700). Patients with dyspnea was 15/120 (12.5%) in the arbidol group and 13/116 (11.2%) in the favipiravir group (P=0.7588). A post-hoc analysis showed that de novo incidences of dyspnea during the course of treatment occurred at 4/116 (3.45%) patients in the favipiravir group and 14/120 (11.67%) patients in the arbidol group (P=0.0174). Hence, we conclude that for secondary outcomes, taking favipiravir significantly shortened the latency to cough relief and reduced fever duration. Favipiravir does not decrease the rate for AOT or NMV, nor does it decrease the overall respiratory failure rate.
Comparison of antiviral-associated adverse effects
During this trial, we detected 37 incidences of antiviral-associated adverse effects (AE) in the favipiravir group and 28 incidences in the arbidol group were observed. All observed AE incidences were level 1. Increased serum uric acid (3 (2.50%) vs 16 (13.79%), P=0.0014) were more common in patients of the favipiravir group than those in the arbidol group. No statistical difference was observed for the frequency of abnormal LFT (ALT and/or AST were elevated, 12 (10.00%) in the arbidol group vs 10 (8.62%) in the favipiravir group; P=0.7156), psychiatric symptom reactions (dysthesia, 1 (0.83%) vs 5 (4.31%); P=0.1149) and digestive tract reactions (nausea, acid reflux, flatulence,11 14 (11.67%) vs 16 (13.79%); P=0.6239) (Table 4). Most of these adverse reactions disappeared by the time patients being discharged.
Comparison of antiviral-associated adverse effects.
The details of ancillary treatments
Besides famiravir and arbidol, the common ancillary treatments included anti-infection, immunomodulator, nutritional support, symptomatic supportive treatment. Ancillary treatments accompanying the experimental drug differ between moderate and severe/critical patients. Details of drugs use were listed in Table S2 and Figure S3. Most frequently used ancillary treatments were: Moxifloxacin Hydrochloride Tablets (22/18, favipiravir/arbidol group); Cephalosporins (11/5); Antiviral drugs other than the experimental drugs (11/27); Glucocorticoid (5/10); and Human Serum Albumin (4/5). Besides above-mentioned medicine, Chinese herbal medicine was widely used among the patients: Lianhua Qingwen Capsule (a prespecified Chinese herbal medicine recipe for respiratory contagious diseases, 23/30); Qiangli Pipa Lu (a syrup for cough relief, 12/15); and Xuebijing Injection (a prespecified Chinese herbal medicine recipe for anti-inflammation, 8/10). We found that whilst additional antiviral treatments were more frequent in the arbidol group (P=0.0045), it did not positively contribute to the clinical recovery rate at day 7.
Discussion
So far, there are no clinically proven effective antiviral drugs for COVID-19. In China, although arbidol has been officially recommended,10 observational experiences suggest that its clinical efficacy and safety were not very optimistic. Efficacy of influenza-directed antiviral agents such as favipiravir for treating COVID-19 remain unclear. We conducted a prospective, multicenter, open-labeled, randomized superiority clinical trial and hypothesized that favipiravir would be superior to arbidol in terms of efficacy for moderate symptoms, and would accelerate the clinical recovery of fever, cough, and breathing difficulties (dyspnea) compared with arbidol. Favipiravir treatment did not improve clinical recovery rate of day 7 (61.21%) compared to arbidol group (51.67%). However, it did significantly improve the latency to cough relief and decreased the duration of fever. Interestingly, post-hoc observation showed that a trend of favipiravir being effective to improve clinical recovery rate in moderate COVID-19 patients (70/98, 71.43%) compared to arbidol (62/111, 55.86%) (P=0.0199). This effect diminished for severe/critical COVID-19 patients. Similarly, post-hoc analysis showed that for moderate COVID-19 patients, auxiliary oxygen therapy or noninvasive mechanical ventilation rate was also decreased in favipiravir group (8/98, 8.16%) compared to arbidol group (19/111, 17.12%) with marginal significance (P=0.0541). Finally, in the FAS, post-hoc analysis also showed that favipiravir treatment significantly decreased de novo incidences of dyspnea (3.45% vs 11.67%, OR=0.27) during the course of treatment. Whether favipiravir would be only effective for moderate COVID-19 patients is a question warrant future investigation. Meanwhile, these results suggest a possibility that favipiravir might be effective for halting disease progress into ARDS, shock and multiple organ failure for moderate COVID-19 patients.
The combination of traditional Chinese medicine and antiviral drugs is more common in China, which is due to the traditional medical culture background of the treatment of choice. Also, anti-infection and immune regulation play an important role in the treatment of the COVID-19. Ancillary treatments, such as traditional Chinese medicine, anti-infection and immunomodulatory drugs, were without statistical difference between groups.
Limitations
For COVID-19, there is no clinically proven effective antiviral drug to serve as the control arm. Although Chinese guideline had recommended several options including recombinant human interferon alpha-2b, ribavirin, chloroquine phosphate, lopinavir and arbidol,10 no RCT results on these drugs were reported. Clinical studies were currently undergoing to test the efficacy and safety of drugs in the treatment of COVID-19. Despite the known antiviral effect of arbidol for influenza and other virus infections, no clinical trial was done to support its effectiveness for SARS-like virus. Arbidol was widely used by Chinese doctors in the beginning stage of this epidemic of COVID-19 (Jan. 1-30, 2020) based on in vitro evidence.12 Hence, for ethical reasons, we chose arbidol as the positive control drug.
Observation time frame was limited due to the urgency of this matter. For the same reason, no relapse (including nucleic acid conversion to positive, fever and cough relapse, or pneumonia progress by radiology) tracking were performed for the discharged patients.
In the inclusion criteria, we did not force positive nucleic acid test as a necessity. The accuracy of nucleic acid assay by throat swab sampling was a known issue among clinical practitioners in China. It was estimated that the assay might have at most 30%-50% of sensitivity for patients in early stage of the disease, whilst contact history, clinical manifestations, radiology evidences, and lab results including leukopenia and lymphopenia were confirmatory for these nucleic-acid-negative pneumonia patients. In the Chinese guideline,10 patients meeting these criteria were considered as confirmed infection. The low sensitivity of PCR-based nucleic acid assay might due to multiple reasons including previous treatment, latency of onset, sampling method, biological specimen characteristics, and the clinical accuracy of specific molecular assay. In this trial, 54/116 patients (46.55%) in favipiravir group and 46/120 (38.33%) in arbidol group were nucleic-acid-positive at day 0. Considering the population incidence of COVID-19 infection at the time of this trial in Wuhan, we consider the probability for mis-identifying patients of pneumonia disease other than COVID-19 into this trial is low.
The eligibility of patients does not prespecify clinical classification as a stratification factor. Ethical concerns arose against completely excluding severe/critical cases from potential beneficial treatment. Additionally, because of the complexity of the disease, progression from moderate to severe/critical is possible. Terminating trial treatment to such patients from the study was considered unacceptable. Post-hoc analysis showed that both treatment and clinical classification contributed significantly to the primary outcome of clinical recovery rate at day 7. Among all participants, there were 18 severe/critical patients in the favipiravir group and 9 severe/critical patients in the arbidol group, which reached a marginal significance (P=0.0531). The difference of clinical classification between groups in the trial had an important impact on its outcome.
Conclusions
Compared to arbidol, favipiravir does not significantly improve clinical recovery rate of day 7. Favipiravir significantly shortened the duration of fever and decreased the latency to cough relief. Antiviral-associated adverse effects of favipiravir is mild and manageable.
Data Availability
With the permission of the corresponding author, we can provide participant data, statistical analysis.
Author contributions
Dr. Wang had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: C. Chen, Huang, Cheng, Wu, Wang.
Acquisition, analysis, or interpretation of data: C. Chen, Huang, Yin, Yi Zhang, Wu, Yongxi Zhang.
Drafting of the manuscript: C. Chen, Huang, Yin, Yi Zhang, S. Chen, Lu, Luo, B. Chen, J. Zhang.
Critical revision of the manuscript for important intellectual content: C. Chen, Huang, Yi Zhang, Wu, S. Chen, Lu, Luo, J. Zhang, B. Chen.
Statistical analysis: C. Chen, Yin, Yi Zhang, S. Chen.
Obtained funding: Wang.
Administrative, technical, or material support: C. Chen, Huang, Wang.
Study supervision: Huang, Yin, Wang.
Declaration of Interest
All authors have no conflicts of interest to declare.
Data sharing Statement
With the permission of the corresponding author, we can provide participant data, statistical analysis.
Role of the Funder/Sponsors
The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Acknowledgments
This work was supported by the National Key Research and Development Program of China (2020YFC0844400).
We would like to acknowledge the support of China National Center for Biotechnology Development, Science and Technology Department of Hubei Province. We thank the excellent technical assistance of Dr. Xiuli Zhao (Beijing Tongren Hospital, Capital Medical University, Beijing, China). We also thank the help from the rescue medical teams for Hubei province (Tianjin medical team, Liaoning medical team, Guangdong Province Traditional Chinese Medical Hospital, Longhua Hospital Shanghai University of Traditional Chinese Medicine, Liyuan Hospital of Tongji Medical College of Huazhong University of Science and Technology, and Puren Hospital of Wuhan University of Science and Technology).
Footnotes
E-mail addresses: changchen{at}whu.edu.cn (Chang Chen), znyylcsy{at}126.com (Jianying Huang), pingyin2000{at}126.com (Ping Yin), zy{at}eulertechnology.com (Yi Zhang), chzs1990{at}163.com(Zhenshun Cheng), wujianyuan{at}znhospital.com (Jianyuan Wu), song.chen{at}whu.edu.cn (Song Chen), znact1936{at}126.com (Yongxi Zhang), znyyssxm{at}126.com (Bo Chen), mxlu_urology{at}whu.edu.cn (Mengxin Lu), luoywen{at}whu.edu.cn (Yongwen Luo), 83351627{at}qq.com (Jingyi Zhang), wangxinghuan{at}whu.edu.cn (Xinghuan Wang).