Abstract
Azvudine and nirmatrelvir-ritonavir (Paxlovid) were widely used to treat patients with COVID-19 in China during the Omicron wave. However, the efficacy and safety of azvudine versus Paxlovid are poorly established. This study included 40,876 hospitalized patients with COVID-19 from eleven hospitals in Henan and Xinjiang Provinces, China. Clinical outcomes were compared between the two drugs via Kaplan-Meier analysis and Cox regression models. Additionally, in vitro and in vivo experiments were used to evaluate the antitumor effects and safety of both drugs. Single-cell RNA sequencing was performed to elucidate the tumor immune landscape after azvudine treatment. After propensity score matching, 2404 azvudine and 1202 Paxlovid recipients from Henan Province were included. Cox regression revealed that azvudine was related to an 18% lower risk of all-cause death than Paxlovid (95% CI: 0.676-0.987), was not obviously different in composite disease progression. The robustness of the findings was verified by the Xinjiang cohort and three sensitivity analyses. Fewer adverse events were observed in the azvudine group. Subgroup analysis revealed that azvudine provided greater benefits for patients with malignant tumors, significantly reducing both all-cause death (hazard ratio [HR]: 0.33, 95% CI: 0.20-0.54) and composite disease progression (HR: 0.54, 95% CI: 0.33-0.88). Furthermore, azvudine can suppress the growth of hepatocellular carcinoma (HCC) by regulating CD4+ T and CD8+ T cells in vivo. These findings suggest that azvudine therapy is not inferior to Paxlovid in hospitalized COVID-19 patients and has fewer adverse effects. Notably, azvudine may offer greater clinical benefit for patients with HCC.