Our results support considering the clinical evaluation of favipiravir against Zika virus but they advocate against its use against SARS-CoV-2 infection. Our models of SARS-CoV-2 infection in reconstituted Human Airway Epithelium (MucilAir) were predictive again of the results obtained in the cynomolgus NHP model, demonstrating that favipiravir, even at doses up to 600 μM, has no antiviral efficacy against SARS-CoV-2. 

Although our results are in contradiction with those obtained in Vero cell lines and hamster models of SARS-CoV-2 infection, they are consistent with the lack of antiviral activity of favipiravir recently reported in randomized clinical trials. This discrepancy clearly highlights the need for a synergistic setup of a pipeline of multiple predictive in vitro and in vivo preclinical models for the robust evaluation of future new antiviral candidates that will maximize their successful implementation in clinical trials.

Conducted by IDMIT, this study was possible thanks to a collaborative network among groups from CEA, l’Inserm, le CNRS – Centre national de la recherche scientifique, l’Institut Pasteur, l’Université de Paris-Saclay, IRD, APHM (Assistance Publique – Hopitaux de Marseille) and Aix-Marseille Université.

The contribution of our team VirPath and its Technology Research Platform VirNext  was supported by Inserm, CNRS – Centre national de la recherche scientifique @INSB CNRS Virocrib and INSTITUT MÉRIEUX.