SummaryBackgroundWell tolerated antivirals administered early in the course of COVID-19 infection when the viremia is highest could prevent progression to severe disease. Favipiravir inhibits SARS-CoV-2 viral replication in vitro with evidence of clinical benefit in open label trials. Placebo controlled studies of people with early symptomatic COVID-19 with regular assessments of SARS-CoV-2 viral load can determine if it has an antiviral effect and improves clinical outcomes.MethodsPeople with PCR-confirmed COVID-19 and 5 days or less of symptoms were randomised 1:1 to favipiravir 1800 mg on day 1, then 800 mg twice daily or matched placebo for 14 days. SARS-CoV-2 viral load was quantitated from second daily self-collected nose-throat swabs while receiving study drug. The primary endpoint was time to virological cure defined as 2 successive swabs negative for SARS-CoV-2 by PCR and secondary outcomes were progression of disease severity, symptom resolution and safety.FindingsBetween 31 July 2020 and 19 September 2021, 200 people were enrolled (199 in the community, 1 in hospital) with 190 receiving one or more doses of drug (modified intention to treat [mITT] population). There was no difference in time to virological cure (Log-rank p=0.6 comparing Kaplan Meier curves), progression to hospitalisation (14 favipiravir, 9 placebo; p=0.38), time to symptom resolution (cough, fever, sore throat) and there were no deaths. 51 people related an adverse event that was possibly drug related, but these were evenly distributed (n=24 favipiravir, n=27 placebo). Sensitivity analyses where the definition of virological cure was changed to: a single negative PCR, exclude datapoints based on the presence or absence of human DNA in the swab, a SARS-CoV-2 viral load < 300 copies/mL being considered negative all demonstrated no difference between arms.InterpretationFavipiravir does not improve the time to virological cure or clinical outcomes and shows no evidence of an antiviral effect when treating early symptomatic COVID-19 infection.FundingThe study was supported in part by grants from the Commonwealth Bank Australia, the Lord Mayor's Charitable Foundation, Melbourne Australia and the Orloff Family Charitable Trust, Melbourne, Australia. JHM is supported by the Medical Research Future Fund, AYP, JT are supported by the Australian National Health and Medical Research Council.
Read full abstract