Abstract
Despite vaccination, many remain vulnerable to coronavirus disease 2019 (COVID-19) and its complications. Oral antivirals to prevent COVID-19 progression are vital. Based on perceived potency and clinical efficacy, favipiravir is widely used to treat COVID-19. Evidence from large randomized controlled trials (RCT) is lacking.
In this multicenter double-blinded placebo-controlled RCT, adults with early mild-to-moderate COVID-19 were 1:1 randomized to favipiravir or placebo. The study evaluated time to sustained clinical recovery (TT-SCR), COVID-19 progression, and cessation of viral shedding.
Of 1187 analyzed patients across 40 centers, 83.3% were Hispanic, 89.0% unvaccinated, 70.3% severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) seronegative, and 77.8% had risk factors for COVID-19 progression. The median time from symptom presentation and from positive test to randomization was 3 and 2 days, respectively. There was no difference in TT-SCR (median of 7 days for both groups; P = .80), COVID-19 progression [11 patients each (1.9% vs 1.8%); P = .96], time to undetectable virus (median = 6 days, 95% confidence interval [CI] [6–8] vs 7 days, 95% CI [6-9]), or in undetectable virus by end of therapy (73.4% vs 72.3%; P = .94). Outcomes were consistent across the analyzed sub-groups. Adverse events were observed in 13.8% and 14.8% of favipiravir-treated and placebo-treated subjects, respectively. Uric acid elevation was more frequent among favipiravir-treated subjects (19.9% vs 2.8%).
Favipiravir was well tolerated but lacked efficacy in TT-SCR, progression to severe COVID-19, or cessation of viral shedding and should not be used to treat patients with COVID-19. (Supported by Appili Therapeutics)
NCT04600895
Prevention remains the mainstay of coronavirus disease 2019 (COVID-19) management, but it is not perfect [1, 2]. Many around the world have limited access to vaccination, do not qualify, or refuse it, resulting in vast unvaccinated populations [3–6]. In addition, the emergence of novel viral variants, which are less affected by vaccines and monoclonal antibodies, combined with the relatively short period of protection and lack of protective immunity among many recipients, results in ongoing transmission, disease, and deaths [7–12]. The availability of effective, safe, and tolerable antivirals to treat early COVID-19 and prevent its progression is therefore of critical importance.
Favipiravir, an RNA-dependent RNA polymerase inhibitor, has been shown to have broad in vitro activity against RNA viruses, including a variety of coronaviruses and specifically severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of COVID-19 [13–15]. Favipiravir was evaluated in multiple studies with results that are widely interpreted as supporting its use in the treatment of patients with COVID-19 [16–21]. This led to its authorization or endorsement in some countries and widespread use outside of the United States [21–25]. However, some trials failed to show a favorable effect or retracted and data from adequately designed double-blinded randomized controlled trials (RCTs) are lacking [26–31]. We present the results of a phase 3 multicenter trial evaluating the efficacy and safety of favipiravir in the treatment of COVID-19.
METHODS
Trial Design and Oversight
The PRESECO (PREvent SEvere COvid-19) study was a randomized, multicenter, double-blinded, placebo-controlled, superiority trial evaluating the safety and efficacy of favipiravir as an outpatient treatment in patients with mild to moderate early COVID-19. This was a 28-day study conducted at 27 sites in the United States, 7 in Brazil, and 6 in Mexico. The study design and choice of endpoints followed the US Food and Drug Administration's (FDA) guidance to industry [32–34]. The study followed Good Clinical Practice principles (ICH Harmonized Tripartite Guideline) [35]. The authors accept responsibility for the accuracy and completeness of the data, the analyses, and adherence to the trial protocol. The study's protocol and statistical analysis plan (SAP) evolved throughout the study to accommodate relevant emerging information on COVID-19 and its management, changing FDA guidance to industry, and updated guidelines from the Centers for Disease Control and Prevention (CDC). Study protocols, SAPs, and description of study amendments are available at NEJM.org.
Trial Participants
Enrollment eligibility included age 18 years or older, mild-to-moderate COVID-19, first positive reverse transcription polymerase chain reaction (RT-PCR) or Rapid Antigen assay within 3 days of enrollment, first symptoms within 5 days of enrollment, having at least 2 moderate or severe COVID-19 symptoms, and signing an informed consent [33, 36]. COVID-19-related symptoms included runny nose, sore throat, shortness of breath, cough, tiredness, body aches, headache, chills, feeling feverish, nausea, diarrhea, and vomiting. Early in the trial, taste or smell changes were included in the entry criteria; however, later, based on an evolving understanding of COVID-19 symptom time course, taste and smell were excluded. Subjects enrolled based solely on taste and smell changes were not analyzed. Exclusion criteria included bacterial or viral infection other than SARS-CoV-2, treatment with high dose steroids, receipt of remdesivir or SARS-CoV-2 monoclonal antibodies. A complete list of inclusion and exclusion criteria is provided in the Appendix.
Efficacy
The primary endpoint of this study was Time to Sustained Clinical Recovery, calculated as the number of days from start of study medication to sustained symptom alleviation, defined by: Oxygen saturation ≥94% at rest, and Oral temperature <38.0°C, and all COVID-19-associated symptoms reaching a score of mild or none for 4 consecutive days [32]. Patients who died, required hospitalization, or withdrew from the study with an unknown outcome were considered not reaching the primary endpoint. A key secondary endpoint was proportion of patients with COVID-19 progression, defined as requiring an emergency department visit or hospitalization for COVID-19 worsening or shortness of breath, or death (“narrow definition”) [32]. An additional definition of progression included the narrow definition criteria, as well as symptomatic worsening, defined as 2 or more additional moderate or severe COVID-19 symptoms, or new fever, or new desaturation (“broad definition”).
Virologic Outcome
An additional prespecified secondary endpoint was time (in days) to undetectable SARS-CoV-2 load in saliva assays. Patients in the viral shedding sub-study collected a saliva specimen daily on study Days 1–10. Patients who had detectable virus at baseline (study Days 1, 2, or 3) were included in the analysis.
Trial Procedures
At enrollment, patients were randomized 1:1 to oral favipiravir 1800 mg twice daily on Day 1, followed by 800 mg twice daily on Days 2–10, or matching placebo for 10 days (Supplementary Figure 1). Participants received the current standard of care. Randomization was performed via an Interactive Web Randomization System and maintained by a third party. Most study interactions were performed by telehealth. To assess COVID-19 manifestations, we followed guidance by the Office of New Drugs at the FDA, which were based upon CDC's recommendations as of 28 August 2020 [33]. Subjects recorded their symptoms, temperature, and O2 saturation, as well as protocol adherence daily, in an electronic patient-reported outcome (ePRO) instrument that was programmed with reminders to complete data entry to reduce missing data. To minimize enrollment bias related to electronic illiteracy, we allowed the use of paper-based patient-reported outcome (PRO) as an alternative [32]. The unblinding code was not broken for any subject throughout the study.
Analysis Populations
The modified-intent-to-treat (mITT) population, which included all randomized patients who received study medication, was the primary analysis population. Additional analysis populations included the Per Protocol (PP) population, which included patients who took at least 80% of study medication and had no protocol violations that would affect the primary efficacy endpoint; and the mITT2 population, which included patients who received study medication and reported (in ePRO) sufficient protocol mandated COVID-19 symptoms (2 or more COVID-19 symptoms worse than mild) on enrollment. The PP and mITT2 populations were used in sensitivity analyses of the primary and key secondary endpoints (Supplementary Table 1). Pre-specified subgroup analyses of the primary and secondary endpoints were performed to assess whether treatment efficacy varied according to serostatus, age, time from positive SARS-CoV-2 test to enrollment, and risk status for COVID-19 progression (defined in the appendix). Additional post hoc analyses included geographical region, race, and uric acid elevation.
Safety
Safety was assessed by documenting adverse events (AEs) and serious AEs (SAEs), according to the Medical Dictionary for Regulatory Activities (MedDRA) system organ class and preferred term (version 23.1) and obtaining laboratory tests.
Statistical Analysis
We planned to enroll 1250 patients to attain 815 clinical recoveries ensuring power of 85%. The definitive sample size calculation was based on group sequential design with a possibility to stop for efficacy or futility at interim. The viral shedding sub-study was powered at 80% with a planned sample size of 500 patients to attain 112 events of viral shedding cessation.
We utilized survival analysis to analyze the time-to-event variables. Failure curves were constructed for time to positive event using Kaplan-Meier estimates. Time to sustained clinical recovery and virologic outcomes were compared using a 2-sided Gehan-Wilcoxon test. Differences between treatments in proportions of COVID-19 progression were evaluated using a 2-sided χ2 test or Fisher exact test, as applicable. The corresponding confidence interval for the difference in proportions were calculated. Adjustment for multiple testing was performed to control the overall type 1 error to ≤5% for the interim and final analysis on the primary efficacy endpoint. A hierarchical approach was also used to control the overall type 1 error to ≤5% for the primary and key secondary endpoints. The order of testing was: (1) Sustained clinical recovery, (2) COVID-19 progression (broad definition), and (3) COVID-19 progression (narrow definition). Statistical significance would not be claimed if the parameter preceding the parameter being tested did not reach statistical significance. Additional information, including handling of missing data, is in the Supplementary Appendix.
RESULTS
Patients
Between November 2020 and October 2021, 1231 patients were randomized; 1187 were included in the mITT population (USA: 963, Mexico: 160, Brazil: 65), 599 received favipiravir, and 588 received placebo (Figure 1). Baseline characteristics were generally similar across treatment groups (Table 1). Most participants were of Hispanic ethnicity (83.3%), unvaccinated (89%), seronegative for COVID-19 nucleocapsid antibodies (70.3%) at enrollment and had risk factors for COVID-19 progression (77.8%). The median number of days from onset of COVID-19 symptoms and from first positive COVID-19 test to randomization was 3 and 2 days, respectively.
Study flow diagram. Abbreviation: COVID-19, coronavirus disease 2019.
Demographics and Baseline Characteristics of the Patients (Modified Intent-to-Treat Population)
| Patient Characteristic | n (%) | |
|---|---|---|
| Placebo(n = 588) | Favipiravir(n = 599) | |
| Male | 261 (44.4%) | 282 (47.1%) |
| Female | 327 (55.6%) | 317 (52.9%) |
| Age group | ||
| ȃ<60 y | 506 (86.1%) | 506 (84.5%) |
| ȃ≥60 y | 82 (13.9%) | 93 (15.5%) |
| Race | ||
| ȃAmerican Indian or Alaska Native | 59 (10.0%) | 63 (10.5%) |
| ȃAsian | 2 (0.3%) | 3 (0.5%) |
| ȃBlack | 26 (4.4%) | 43 (7.2%) |
| ȃNative Hawaiian or Pacific Islander | 1 (0.2%) | 0 |
| ȃWhite | 491 (83.5%) | 485 (81.0%) |
| ȃOther or not reported | 9 (1.5%) | 5 (0.8%) |
| Ethnicity | ||
| ȃHispanic or Latino | 495 (84.2%) | 494 (82.5%) |
| ȃNon-Hispanic or Latino | 93 (15.8%) | 102 (17.0%) |
| ȃNot Reported or Unknown | 0 | 3 (0.5%) |
| Country | ||
| ȃUnited States | 480 (81.6%) | 482 (80.5%) |
| ȃMexico | 78 (13.3%) | 82 (13.7%) |
| ȃBrazil | 30 (5.1%) | 35 (5.8%) |
| Vaccinated for COVID-19 at Enrollment | 72 (12.2%) | 59 (9.8%) |
| Seropositive at enrollment | 173 (29.4%) | 159 (26.5%) |
| Risk status at enrollment for COVID-19 progressiona | ||
| ȃHigh risk | 453 (77.0%) | 470 (78.5%) |
| ȃȃAge ≥65 y | 44 (7.5%) | 56 (9.3%) |
| ȃȃBMI (kg/m2) ≥25 | 420 (71.4%) | 435 (72.6%) |
| ȃȃAny qualifying comorbidity (at least 1 condition #4–12)a | 151 (25.7%) | 62 (27.0%) |
| ȃȃMore than 1 risk factor or comorbidity (#1–12)a | 145 (24.7%) | 152 (25.4%) |
| ȃNot high risk | 135 (23.0%) | 129 (21.5%) |
| COVID-19 symptom severity at enrollment | ||
| ȃMultiple moderate or severe symptoms (≥4) | 306 (52.0%) | 293 (48.9%) |
| ȃFew moderate or severe symptoms (2–3) | 281 (47.8%) | 305 (50.9%) |
| Days from first positive SARS-CoV-2 test to randomizationb | 2 (0,3) | 2 (0,3) |
| ȃ≤2 d | 521 (88.6%) | 528 (88.1%) |
| ȃ2–5 d | 66 (11.2%) | 71 (11.9%) |
| No. of days from onset of COVID-19 symptoms to randomizationb | 3 (0,7) | 3 (0,6) |
| Patient Characteristic | n (%) | |
|---|---|---|
| Placebo(n = 588) | Favipiravir(n = 599) | |
| Male | 261 (44.4%) | 282 (47.1%) |
| Female | 327 (55.6%) | 317 (52.9%) |
| Age group | ||
| ȃ<60 y | 506 (86.1%) | 506 (84.5%) |
| ȃ≥60 y | 82 (13.9%) | 93 (15.5%) |
| Race | ||
| ȃAmerican Indian or Alaska Native | 59 (10.0%) | 63 (10.5%) |
| ȃAsian | 2 (0.3%) | 3 (0.5%) |
| ȃBlack | 26 (4.4%) | 43 (7.2%) |
| ȃNative Hawaiian or Pacific Islander | 1 (0.2%) | 0 |
| ȃWhite | 491 (83.5%) | 485 (81.0%) |
| ȃOther or not reported | 9 (1.5%) | 5 (0.8%) |
| Ethnicity | ||
| ȃHispanic or Latino | 495 (84.2%) | 494 (82.5%) |
| ȃNon-Hispanic or Latino | 93 (15.8%) | 102 (17.0%) |
| ȃNot Reported or Unknown | 0 | 3 (0.5%) |
| Country | ||
| ȃUnited States | 480 (81.6%) | 482 (80.5%) |
| ȃMexico | 78 (13.3%) | 82 (13.7%) |
| ȃBrazil | 30 (5.1%) | 35 (5.8%) |
| Vaccinated for COVID-19 at Enrollment | 72 (12.2%) | 59 (9.8%) |
| Seropositive at enrollment | 173 (29.4%) | 159 (26.5%) |
| Risk status at enrollment for COVID-19 progressiona | ||
| ȃHigh risk | 453 (77.0%) | 470 (78.5%) |
| ȃȃAge ≥65 y | 44 (7.5%) | 56 (9.3%) |
| ȃȃBMI (kg/m2) ≥25 | 420 (71.4%) | 435 (72.6%) |
| ȃȃAny qualifying comorbidity (at least 1 condition #4–12)a | 151 (25.7%) | 62 (27.0%) |
| ȃȃMore than 1 risk factor or comorbidity (#1–12)a | 145 (24.7%) | 152 (25.4%) |
| ȃNot high risk | 135 (23.0%) | 129 (21.5%) |
| COVID-19 symptom severity at enrollment | ||
| ȃMultiple moderate or severe symptoms (≥4) | 306 (52.0%) | 293 (48.9%) |
| ȃFew moderate or severe symptoms (2–3) | 281 (47.8%) | 305 (50.9%) |
| Days from first positive SARS-CoV-2 test to randomizationb | 2 (0,3) | 2 (0,3) |
| ȃ≤2 d | 521 (88.6%) | 528 (88.1%) |
| ȃ2–5 d | 66 (11.2%) | 71 (11.9%) |
| No. of days from onset of COVID-19 symptoms to randomizationb | 3 (0,7) | 3 (0,6) |
Abbreviations: BMI, body mass index; COVID-19, coronavirus disease 2019; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
High risk for COVID-19 progression is defined as meeting at least 1 of the following: (1) 65 years of age or older OR (2) Overweight (BMI ≥25) OR (3) Pregnancy OR (4) Chronic kidney disease OR (5) Diabetes OR (6) Weakened immune system OR (7) Currently receiving immunosuppressive treatment OR (8) Cardiovascular disease/hypertension OR (9) Chronic lung disease OR (10) Sickle cell disease OR (11) Neurodevelopmental disorders OR (12) Medical-related technological dependence.
Median (min, max).
Demographics and Baseline Characteristics of the Patients (Modified Intent-to-Treat Population)
| Patient Characteristic | n (%) | |
|---|---|---|
| Placebo(n = 588) | Favipiravir(n = 599) | |
| Male | 261 (44.4%) | 282 (47.1%) |
| Female | 327 (55.6%) | 317 (52.9%) |
| Age group | ||
| ȃ<60 y | 506 (86.1%) | 506 (84.5%) |
| ȃ≥60 y | 82 (13.9%) | 93 (15.5%) |
| Race | ||
| ȃAmerican Indian or Alaska Native | 59 (10.0%) | 63 (10.5%) |
| ȃAsian | 2 (0.3%) | 3 (0.5%) |
| ȃBlack | 26 (4.4%) | 43 (7.2%) |
| ȃNative Hawaiian or Pacific Islander | 1 (0.2%) | 0 |
| ȃWhite | 491 (83.5%) | 485 (81.0%) |
| ȃOther or not reported | 9 (1.5%) | 5 (0.8%) |
| Ethnicity | ||
| ȃHispanic or Latino | 495 (84.2%) | 494 (82.5%) |
| ȃNon-Hispanic or Latino | 93 (15.8%) | 102 (17.0%) |
| ȃNot Reported or Unknown | 0 | 3 (0.5%) |
| Country | ||
| ȃUnited States | 480 (81.6%) | 482 (80.5%) |
| ȃMexico | 78 (13.3%) | 82 (13.7%) |
| ȃBrazil | 30 (5.1%) | 35 (5.8%) |
| Vaccinated for COVID-19 at Enrollment | 72 (12.2%) | 59 (9.8%) |
| Seropositive at enrollment | 173 (29.4%) | 159 (26.5%) |
| Risk status at enrollment for COVID-19 progressiona | ||
| ȃHigh risk | 453 (77.0%) | 470 (78.5%) |
| ȃȃAge ≥65 y | 44 (7.5%) | 56 (9.3%) |
| ȃȃBMI (kg/m2) ≥25 | 420 (71.4%) | 435 (72.6%) |
| ȃȃAny qualifying comorbidity (at least 1 condition #4–12)a | 151 (25.7%) | 62 (27.0%) |
| ȃȃMore than 1 risk factor or comorbidity (#1–12)a | 145 (24.7%) | 152 (25.4%) |
| ȃNot high risk | 135 (23.0%) | 129 (21.5%) |
| COVID-19 symptom severity at enrollment | ||
| ȃMultiple moderate or severe symptoms (≥4) | 306 (52.0%) | 293 (48.9%) |
| ȃFew moderate or severe symptoms (2–3) | 281 (47.8%) | 305 (50.9%) |
| Days from first positive SARS-CoV-2 test to randomizationb | 2 (0,3) | 2 (0,3) |
| ȃ≤2 d | 521 (88.6%) | 528 (88.1%) |
| ȃ2–5 d | 66 (11.2%) | 71 (11.9%) |
| No. of days from onset of COVID-19 symptoms to randomizationb | 3 (0,7) | 3 (0,6) |
| Patient Characteristic | n (%) | |
|---|---|---|
| Placebo(n = 588) | Favipiravir(n = 599) | |
| Male | 261 (44.4%) | 282 (47.1%) |
| Female | 327 (55.6%) | 317 (52.9%) |
| Age group | ||
| ȃ<60 y | 506 (86.1%) | 506 (84.5%) |
| ȃ≥60 y | 82 (13.9%) | 93 (15.5%) |
| Race | ||
| ȃAmerican Indian or Alaska Native | 59 (10.0%) | 63 (10.5%) |
| ȃAsian | 2 (0.3%) | 3 (0.5%) |
| ȃBlack | 26 (4.4%) | 43 (7.2%) |
| ȃNative Hawaiian or Pacific Islander | 1 (0.2%) | 0 |
| ȃWhite | 491 (83.5%) | 485 (81.0%) |
| ȃOther or not reported | 9 (1.5%) | 5 (0.8%) |
| Ethnicity | ||
| ȃHispanic or Latino | 495 (84.2%) | 494 (82.5%) |
| ȃNon-Hispanic or Latino | 93 (15.8%) | 102 (17.0%) |
| ȃNot Reported or Unknown | 0 | 3 (0.5%) |
| Country | ||
| ȃUnited States | 480 (81.6%) | 482 (80.5%) |
| ȃMexico | 78 (13.3%) | 82 (13.7%) |
| ȃBrazil | 30 (5.1%) | 35 (5.8%) |
| Vaccinated for COVID-19 at Enrollment | 72 (12.2%) | 59 (9.8%) |
| Seropositive at enrollment | 173 (29.4%) | 159 (26.5%) |
| Risk status at enrollment for COVID-19 progressiona | ||
| ȃHigh risk | 453 (77.0%) | 470 (78.5%) |
| ȃȃAge ≥65 y | 44 (7.5%) | 56 (9.3%) |
| ȃȃBMI (kg/m2) ≥25 | 420 (71.4%) | 435 (72.6%) |
| ȃȃAny qualifying comorbidity (at least 1 condition #4–12)a | 151 (25.7%) | 62 (27.0%) |
| ȃȃMore than 1 risk factor or comorbidity (#1–12)a | 145 (24.7%) | 152 (25.4%) |
| ȃNot high risk | 135 (23.0%) | 129 (21.5%) |
| COVID-19 symptom severity at enrollment | ||
| ȃMultiple moderate or severe symptoms (≥4) | 306 (52.0%) | 293 (48.9%) |
| ȃFew moderate or severe symptoms (2–3) | 281 (47.8%) | 305 (50.9%) |
| Days from first positive SARS-CoV-2 test to randomizationb | 2 (0,3) | 2 (0,3) |
| ȃ≤2 d | 521 (88.6%) | 528 (88.1%) |
| ȃ2–5 d | 66 (11.2%) | 71 (11.9%) |
| No. of days from onset of COVID-19 symptoms to randomizationb | 3 (0,7) | 3 (0,6) |
Abbreviations: BMI, body mass index; COVID-19, coronavirus disease 2019; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
High risk for COVID-19 progression is defined as meeting at least 1 of the following: (1) 65 years of age or older OR (2) Overweight (BMI ≥25) OR (3) Pregnancy OR (4) Chronic kidney disease OR (5) Diabetes OR (6) Weakened immune system OR (7) Currently receiving immunosuppressive treatment OR (8) Cardiovascular disease/hypertension OR (9) Chronic lung disease OR (10) Sickle cell disease OR (11) Neurodevelopmental disorders OR (12) Medical-related technological dependence.
Median (min, max).
Efficacy
The median time to sustained clinical recovery in the favipiravir and placebo groups, was 7 days (95% confidence interval [CI] 7–8 days) and 7 days (95% CI 6–8), respectively. By study Day 28, sustained clinical recovery was reached by 87.8% of favipiravir-treated and 87.3% placebo-treated subjects (P = .80) (Figure 2A). Because the primary analysis showed no statistical significance, the hierarchical analyses approach mandated that all secondary efficacy analyses are considered as not having reached statistical significance. Regardless, analyses of the key secondary endpoints were performed and failed to reach statistical significance. In each the favipiravir and placebo groups, 11 patients (1.9% vs 1.8%) experienced COVID-19 progression (narrow definition), (P = .96) (Figure 2B). Similarly, there was no difference in COVID-19 progression using the broad definition for progression between favipiravir-treated and placebo-treated patients (Supplementary Table 2). Outcomes were consistent across all analysis populations and a lack of efficacy of favipiravir in relation to the primary and key secondary endpoints was seen in the additional analysis populations (Supplementary Figure 2). A lack of efficacy was also observed across all treatment groups in the analyzed subgroups for the primary and secondary endpoints.
Efficacy of favipiravir as an outpatient treatment for patients with early mild to moderate COVID-19 (modified intent-to-treat population). A, Time to sustained clinical recovery, favipiravir versus placebo. B, Sustained clinical recovery and progression to severe COVID-19 overall and stratified by patient's characteristics. Abbreviations: CI, confidence interval; COVID-19, coronavirus disease 2019; EOT, end of therapy; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
Virologic Outcome
Of the 482 patients in the mITT viral shedding sub-study, 272 (favipiravir 140, placebo 132) had detectable viral load at baseline (study Day 1, 2, or 3) and were included in the analysis. Approximately 50% of patients had converted to undetectable SARS-CoV-2 viral load by Day 6 (95% CI 6–8 days) in the favipiravir group and by Day 7 (95% CI 6–8 days) in the placebo group; 73.4% of patients in the favipiravir treatment group reached undetectable SARS-CoV-2 Viral load by Day 10, compared with 72.3% in the placebo group. There was no difference in time to cessation of viral shedding between the treatment groups (P = .94) (Figure 3). Outcomes were consistent across analysis populations (Supplementary Figure 3).
Time to negative conversion of detectable SARS-CoV-2 viral RNA in RT-PCR assays of saliva (modified intent-to-treat population). P = .94. Abbreviations: RT-PCR, reverse-transcription polymerase chain reaction; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
Safety
Adverse events occurred in 84 patients (13.8%) in the favipiravir group and 89 patients (14.8%) in the placebo group. Of these, 14 (2.3%) in the favipiravir group and 20 (3.3%) in the placebo group were considered possibly, probably, or definitely related to study drug (Table 2 and Supplementary Table 3). Treatment discontinuations due to AEs occurred in 9 patients (1.5%) and 7 patients (1.2%) in the favipiravir and placebo groups, respectively. The most common AEs included nausea (favipiravir: 8 patients, 1.3% vs placebo: 5 patients, 0.8%), cough (7 patients, 1.1% vs 5 patients, 0.8%), and dizziness (5 patients, 0.8% vs 10 patients, 1.7%) (Table 2). SAEs occurred in 12 (2.0%) and 14 (2.3%) patients in the favipiravir and placebo groups, respectively (Table 2 and Supplementary Table 4). The most common SAEs included COVID-19 pneumonia in 5 (0.8%) patients in each group, pneumonia (favipiravir: 3 patients, 0.5% vs placebo: 1 patient, 0.2%), COVID-19 worsening (2 patients, 0.3% vs 3 patients, 0.5%), and dyspnea (1 patient, 0.2% vs 2 patients, 0.3%). Study investigators determined all SAEs in the favipiravir group unrelated to study treatment. One death occurred during the study in a patient receiving placebo. The death was attributed to complications of COVID-19 pneumonia (Table 2). Abnormally elevated uric acid levels at treatment end were seen in 105 favipiravir-treated patients (19.9%, 95% CI 16.5%–23.3%) versus 15 placebo-treated (2.8%, 95% CI 1.4%–4.3%), and the mean uric acid change in favipiravir-treated patients was 30.1% (standard deviation [SD] 51.6) versus 6.6% (SD 21.2) in placebo-treated patients (Table 2). Abnormally elevated alanine aminotransferase at treatment end was seen in 68 favipiravir-treated (12.9%, 95% CI 10.1%-15.8%) versus 44 placebo-treated (8.3%, 95% CI 6.0%–10.7%) patients (Supplementary Table 5).
Summary of Adverse Events (Safety Population)
| Adverse Event Category | n (%) [no. of events] | |
|---|---|---|
| Placebo (n = 601) | Favipiravir (n = 610) | |
| Any adverse event | 89 (14.8%) [160] | 84 (13.8%) [149] |
| Serious adverse event | 14 (2.3%) [17] | 12 (2.0%) [16] |
| CTCAE Grade 3 or 4 TEAE | 17 (2.8%) [18] | 14 (2.3%) [18] |
| Related adverse event | 20 (3.3%) [25] | 14 (2.3%) [18] |
| Related serious adverse event | 2 (0.3%) [2] | 0 |
| Related CTCAE Grade 3 or 4 TEAE | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Adverse event leading to treatment discontinuation | 7 (1.2%) [10] | 9 (1.5%) [14] |
| Adverse event leading to study discontinuation | 9 (1.5%) [11] | 7 (1.1%) [9] |
| Adverse event leading to deatha | 1 (0.2%) [4] | 0 |
| Most common adverse eventsb | ||
| ȃNausea | 5 (0.8%) [7] | 8 (1.3%) [9] |
| ȃCough | 5 (0.8%) [5] | 7 (1.1%) [7] |
| ȃDizziness | 10 (1.7%) [10] | 5 (0.8%) [5] |
| Most common serious adverse eventsc | ||
| ȃCOVID-19 pneumonia | 5 (0.8%) [5] | 5 (0.8%) [5] |
| ȃPneumonia | 1 (0.2%) [1] | 3 (0.5%) [3] |
| ȃCOVID-19 worsening | 3 (0.5%) [3] | 2 (0.3%) [2] |
| ȃDyspnea | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Blood uric acid % increase from baselined | 6.6% (21.2) | 30.1% (51.6) |
| Alanine aminotransferase % increase from baselined | 16.4% (94.6) | 27.3% (95.0) |
| Aspartate aminotransferase % increase from baselined | −1.1% (41.1) | 9.3% (93.2) |
| Adverse Event Category | n (%) [no. of events] | |
|---|---|---|
| Placebo (n = 601) | Favipiravir (n = 610) | |
| Any adverse event | 89 (14.8%) [160] | 84 (13.8%) [149] |
| Serious adverse event | 14 (2.3%) [17] | 12 (2.0%) [16] |
| CTCAE Grade 3 or 4 TEAE | 17 (2.8%) [18] | 14 (2.3%) [18] |
| Related adverse event | 20 (3.3%) [25] | 14 (2.3%) [18] |
| Related serious adverse event | 2 (0.3%) [2] | 0 |
| Related CTCAE Grade 3 or 4 TEAE | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Adverse event leading to treatment discontinuation | 7 (1.2%) [10] | 9 (1.5%) [14] |
| Adverse event leading to study discontinuation | 9 (1.5%) [11] | 7 (1.1%) [9] |
| Adverse event leading to deatha | 1 (0.2%) [4] | 0 |
| Most common adverse eventsb | ||
| ȃNausea | 5 (0.8%) [7] | 8 (1.3%) [9] |
| ȃCough | 5 (0.8%) [5] | 7 (1.1%) [7] |
| ȃDizziness | 10 (1.7%) [10] | 5 (0.8%) [5] |
| Most common serious adverse eventsc | ||
| ȃCOVID-19 pneumonia | 5 (0.8%) [5] | 5 (0.8%) [5] |
| ȃPneumonia | 1 (0.2%) [1] | 3 (0.5%) [3] |
| ȃCOVID-19 worsening | 3 (0.5%) [3] | 2 (0.3%) [2] |
| ȃDyspnea | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Blood uric acid % increase from baselined | 6.6% (21.2) | 30.1% (51.6) |
| Alanine aminotransferase % increase from baselined | 16.4% (94.6) | 27.3% (95.0) |
| Aspartate aminotransferase % increase from baselined | −1.1% (41.1) | 9.3% (93.2) |
Abbreviations: COVID-19, coronavirus disease 2019; CTCAE, common terminology criteria for adverse events; TEAE, treatment emergent adverse events.
1 subject in the placebo group died as a result of COVID-19 pneumonia, acute respiratory failure, pneumomediastinum, and hypoxia.
Reported in ≥1% of patients, overall.
Reported in ≥0.2% of patients, overall.
Mean (SD).
Summary of Adverse Events (Safety Population)
| Adverse Event Category | n (%) [no. of events] | |
|---|---|---|
| Placebo (n = 601) | Favipiravir (n = 610) | |
| Any adverse event | 89 (14.8%) [160] | 84 (13.8%) [149] |
| Serious adverse event | 14 (2.3%) [17] | 12 (2.0%) [16] |
| CTCAE Grade 3 or 4 TEAE | 17 (2.8%) [18] | 14 (2.3%) [18] |
| Related adverse event | 20 (3.3%) [25] | 14 (2.3%) [18] |
| Related serious adverse event | 2 (0.3%) [2] | 0 |
| Related CTCAE Grade 3 or 4 TEAE | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Adverse event leading to treatment discontinuation | 7 (1.2%) [10] | 9 (1.5%) [14] |
| Adverse event leading to study discontinuation | 9 (1.5%) [11] | 7 (1.1%) [9] |
| Adverse event leading to deatha | 1 (0.2%) [4] | 0 |
| Most common adverse eventsb | ||
| ȃNausea | 5 (0.8%) [7] | 8 (1.3%) [9] |
| ȃCough | 5 (0.8%) [5] | 7 (1.1%) [7] |
| ȃDizziness | 10 (1.7%) [10] | 5 (0.8%) [5] |
| Most common serious adverse eventsc | ||
| ȃCOVID-19 pneumonia | 5 (0.8%) [5] | 5 (0.8%) [5] |
| ȃPneumonia | 1 (0.2%) [1] | 3 (0.5%) [3] |
| ȃCOVID-19 worsening | 3 (0.5%) [3] | 2 (0.3%) [2] |
| ȃDyspnea | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Blood uric acid % increase from baselined | 6.6% (21.2) | 30.1% (51.6) |
| Alanine aminotransferase % increase from baselined | 16.4% (94.6) | 27.3% (95.0) |
| Aspartate aminotransferase % increase from baselined | −1.1% (41.1) | 9.3% (93.2) |
| Adverse Event Category | n (%) [no. of events] | |
|---|---|---|
| Placebo (n = 601) | Favipiravir (n = 610) | |
| Any adverse event | 89 (14.8%) [160] | 84 (13.8%) [149] |
| Serious adverse event | 14 (2.3%) [17] | 12 (2.0%) [16] |
| CTCAE Grade 3 or 4 TEAE | 17 (2.8%) [18] | 14 (2.3%) [18] |
| Related adverse event | 20 (3.3%) [25] | 14 (2.3%) [18] |
| Related serious adverse event | 2 (0.3%) [2] | 0 |
| Related CTCAE Grade 3 or 4 TEAE | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Adverse event leading to treatment discontinuation | 7 (1.2%) [10] | 9 (1.5%) [14] |
| Adverse event leading to study discontinuation | 9 (1.5%) [11] | 7 (1.1%) [9] |
| Adverse event leading to deatha | 1 (0.2%) [4] | 0 |
| Most common adverse eventsb | ||
| ȃNausea | 5 (0.8%) [7] | 8 (1.3%) [9] |
| ȃCough | 5 (0.8%) [5] | 7 (1.1%) [7] |
| ȃDizziness | 10 (1.7%) [10] | 5 (0.8%) [5] |
| Most common serious adverse eventsc | ||
| ȃCOVID-19 pneumonia | 5 (0.8%) [5] | 5 (0.8%) [5] |
| ȃPneumonia | 1 (0.2%) [1] | 3 (0.5%) [3] |
| ȃCOVID-19 worsening | 3 (0.5%) [3] | 2 (0.3%) [2] |
| ȃDyspnea | 2 (0.3%) [2] | 1 (0.2%) [1] |
| Blood uric acid % increase from baselined | 6.6% (21.2) | 30.1% (51.6) |
| Alanine aminotransferase % increase from baselined | 16.4% (94.6) | 27.3% (95.0) |
| Aspartate aminotransferase % increase from baselined | −1.1% (41.1) | 9.3% (93.2) |
Abbreviations: COVID-19, coronavirus disease 2019; CTCAE, common terminology criteria for adverse events; TEAE, treatment emergent adverse events.
1 subject in the placebo group died as a result of COVID-19 pneumonia, acute respiratory failure, pneumomediastinum, and hypoxia.
Reported in ≥1% of patients, overall.
Reported in ≥0.2% of patients, overall.
Mean (SD).
DISCUSSION
In this multicenter, double-blind RCT, we evaluated the efficacy and safety of favipiravir in patients with mild to moderate early COVID-19. Favipiravir treatment was associated with uric acid elevation but no documented episodes of gout. Otherwise, favipiravir was safe and well tolerated. Favipiravir lacked a mitigating effect on COVID-19 symptoms, measured as time to sustained clinical recovery, and did not decrease the progression of COVID-19, measured as the development of new moderate or severe symptoms or the need for emergency room visits and hospitalizations. In addition, the pace to cessation of viral shedding was not affected by favipiravir treatment. This lack of favipiravir effect was observed across analysis populations and patient subgroups.
Several factors may explain favipiravir's lack of efficacy. Insufficient potency is a leading concern. Favipiravir was not systematically studied in vitro against most SARS-CoV-2 variants. The amount of information for variants that have been studied is limited and demonstrates a high ED50, suggesting that higher drug concentrations than may be achievable with the studied dose are required for efficacy [14, 15]. COVID-19 animal studies, as well as pharmacologic modeling to fully understand the pharmacokinetic-pharmacodynamic profile of favipiravir, inform dosing decisions, and mitigate subtherapeutic exposures, are not available. In addition, it is not clear whether favipiravir exposure may vary across populations of patients with COVID-19, requiring better understanding of dosing considerations among patients who are non-Asian, young, or overweight. Although the patients’ self-documented medication adherence was high, undocumented lack of medication adherence related to high pill burden could have contributed to inconsistent exposure. Finally, some patients may have been further into their COVID-19 course and therefore out of the essential viral-replication window for antivirals to exert a beneficial effect.
Results of studies evaluating favipiravir's efficacy, published and unpublished, are mixed and possibly misleading by supporting a favorable effect [16–20, 26–31]. Such studies can be divided into anecdotal case series and quasi-experimental studies, all of which were open-label and mostly small in sample size [16–20]. Only a few controlled double-blind studies have been performed; however, they were underpowered, and some were terminated early for futility [26–31]. A telling trend emerges when reviewing the results. Many of the open-label quasi-experimental studies demonstrated efficacy of favipiravir in mitigating COVID-19 symptoms or improving outcomes [16–20]. However, of the 6 RCTs conducted, only 1 was double-blind, none demonstrated proven clinical efficacy, and none demonstrated effective viral cessation [26–31]. Positive sentiments emerging from open-label quasi-experimental studies have been the basis for the widespread endorsement, even regulatory authorization in some countries, of using favipiravir for the management of COVID-19 [22–25].
The PRESECO study was limited in several ways. A major challenge in performing trials early in the pandemic is the rapidly evolving understanding of the disease, its manifestations and course, as well as frequent changes in guidelines for its diagnosis and management. Although protocol changes during a study are generally not endorsed, unless a study evolves as relevant information becomes available along a pandemic, its results may be deemed irrelevant by the end of it. We avoided major changes to the study's design and its endpoints, while updating the sample size to achieve the prespecified number of events and therefore maintaining study's objectives and integrity. The primary endpoint of this study is subjective and difficult to measure. To overcome this challenge, we used PRO with a 4-point scale for severity designed to be completed in real time in an electronic diary by the patient [33]. This validated standardized approach and the large sample size helped reduce documentation bias. Another challenge was ensuring adequate representation of all affected by COVID-19 regardless of their ability to use an electronic diary. To address this challenge, we allowed the use of a paper diary with frequent reminders from the study team. In this study, the proportion of patients with COVID-19 progression was relatively small. Yet the complete lack of favipiravir efficacy on progression in this study suggests that a larger sample size would have been unlikely to show a favorable effect.
In summary, adequately designed clinical trials are important, particularly during pandemic times, when favorable study results can be immediately translated to better care and outcomes. Conversely, inconclusively designed trials can produce misleading results and drain the pool of study candidates for the meaningful trials. Contradicting much of the data from quasi-experimental studies and despite signals of in vitro potency, in this multicenter RCT, favipiravir showed no evidence of clinical efficacy or viral shedding cessation. Until data from additional adequately designed studies using different regimens of favipiravir formulations or dosing becomes available, we advise against the use of favipiravir in the treatment of COVID-19.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments. The authors would like to thank Prasann Bavaniya and Lussia Kim for their critical contribution to the conduct of this clinical trial.
Financial support. The conduct of this clinical trial and the preparation of this manuscript were supported by Appili Therapeutics, Inc., Halifax, NS, Canada.
References
Author notes
Potential Conflicts of Interest. Y. G. is an employee of Appili Therapeutics in the capacity of Chief Medical Officer. Speaker: Pfizer, Sanofi, AbbVie, Paratek, Shionogi, Merk. Consultant/advisor: Pfizer, Seres, Sanofi. Stock options: Appili Therapeutics, Inc. Grants from NIH, CO-PI, niclosamide in COVID-19. J. A. S. C. is Advisor and speaker: Pfizer, Astra Zeneca, and Regeneron (paid to author). Served as site PI in this clinical trial and reimbursed for subject enrollment. Consulting fees include advisory board Astrazeneca and advisory board Pfizer (paid to author). R. W. is an employee of Rho, Inc, Durham, North Carolina, USA. Served as the study's statistician. D. C. is an employee of Appili Therapeutics in the capacity of Chief Development Officer. Stock options: Appili Therapeutics, Inc. R. H. served as site PI in this clinical trial and reimbursed for subject enrollment. Y. G.-R. served as site PI in this clinical trial and reimbursed for subject enrollment. R. L. served as site PI in this clinical trial and reimbursed for subject enrollment. R. F. (deceased) served as site PI in this clinical trial and reimbursed for subject enrollment. A. B. is an employee of Appili Therapeutics, Inc, in the capacity of Chief Executive Officer. Stock options: Appili Therapeutics, Inc; includes grant or contracts from US DoD – Research Grants unrelated to COVID-19/Favipiravir. All other authors report no potential conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
