Patients infected with Omicron subvariant BA.2.76 or BA.5.1 were retrospectively recruited into this study from the Fifth Hospital of Shijiazhuang [designated hospital for patients with Coronavirus disease 2019 (COVID-19)] between August 10, 2022, and October 9, 2022. The inclusion criteria for participants as follows: (1) aged 18 years or older; (2) positive for SARS-CoV-2 by real-time reverse transcriptase-polymerase chain reaction (RT-PCR) from nasopharyngeal and/or oropharyngeal swabs; (3) source patients of the epidemic outbreak were diagnosed with Omicron BA.2.76 or BA.5.1 infection by gene sequencing; (4) no previous infection with other SARS-CoV-2 strains from self-report and medical records; (5) was transferred to the Fifth Hospital of Shijiazhuang according to the requirements of the COVID-19 prevention and control protocol¹ and (6) willing to participate in the study and sign the informed consent. Exclusion criteria were as follows: (1) patients with not detailed medical records and laboratory examination results; (2) patients receiving Azvudine treatment for HIV; (3) pregnant women or lactating mothers. A total of 244 patients were included in this study. We divided the patients into two groups according to the Omicron strain: BA.2.76 group (n = 158) and BA.5.1 group (n = 86).

The clinical characteristics of the enrolled patients, including COVID-19 symptoms, underlying diseases, and vaccination status, were obtained from the hospital information system at admission and collected before treatment. According to the 9^th edition of COVID-19 protocols for diagnosis and treatment (19), the patients recruited in the present study had non-severe (asymptomatic, mild, and moderate) illness. All participants had a clear clinical diagnosis of infection severity on the first day after admission. Nucleic acid testing was performed every one to two days for each patient during hospitalization.

Based on clinical symptoms, nucleic acid cycle threshold (CT), and lung infection, 121 patients in the study received Azvudine drug (5 mg/day for 7–14 days, Henan Genuine Biotech) and two types of self-developed Chinese medicines, including Qingre Kangdu or Lanxiang Jiedu herbal extracts (once daily). The patients in Azvudine group received Azivudine from the second (interquartile range: 1, 2) day after they tested positive for SARS-CoV-2, and the course of treatment lasted a maximum of 14 days. The non-Azvudine treatment group comprised 123 patients who only received either Qingre Kangdu or Lanxiang Jiedu herbal extracts (once daily) during hospitalization.

All patients provided informed consent, and the Institutional Review Board of the Fifth Hospital of Shijiazhuang approved this study (2022001).

The data on routine laboratory indices that reflect inflammation, coagulation, blood cell parameters and immunology were derived from the laboratory information system. C-reactive protein (CRP), procalcitonin (PCT), interleukin (IL), and other indexes were performed on Canon TOSHIBA-FX8 biochemistry analyzer. Blood cell counts were determined using Sysmex XN-1000 Pure Hematology Analyzer. Levels of anti-SARS-CoV-2 IgG and IgM were evaluated using YHLO iFlash 3,000 immunoassay analyzer. RT-PCR for SARS-CoV-2 was performed on Applied Biosystems 7,500 Real-Time PCR system. Except for IL, which was not assessed before discharge, all other indicators were detected on the first day of admission after infection (before treatment) and the day before discharge.

Statistical analysis was performed using R version 4.1.3 software, IBM SPSS Statistics version 26.0 (IBM Corp, USA), and Prism 8.0 (GraphPad, San Diego, California, USA). Normality testing was conducted using the Kolmogorov–Smirnov test. Quantitative data with a normal or non-normal distribution were expressed as mean (standard deviation, SD) and median (interquartile range, IQR), respectively. Categorical variables were presented as numbers and percentages. Independent sample t-test and Wilcoxon rank-sum test were, respectively, applied to analyze normally and non-normally distributed data. For categorical variables, the χ² test was performed.

Binary logistic regression analysis was performed using SPSS software. Taking the median NAN durations of (10 days) patients in this study as threshold, binary logistic regression was initially performed to define the clinical (Azvudine treatment, BA.5.1 subvariant, fever symptom) and laboratory (lymphocyte counts and anti-SARS-CoV-2 IgG before treatment) risk factors associated with the prolonged virus clearance (>10 days) of patients infected with Omicron subvariant. The models included the adjusted Model 1 (intaking clinical parameters) and adjusted Model 2 (intaking clinical and laboratory factors).

Besides, indexes with statistical differences between patients with and without pneumonia were included in a univariable analysis, which were used to calculate odds ratios (OR) and 95% confidence intervals (CI) to explore factors associated with the incidence of pneumonia. The following factors were evaluated in the univariable analyses: fatigue (yes vs. no), dyspnea (yes vs. no), and levels of D-dimer. Moreover, statistically significant variables associated with pneumonia in the univariable analyses were further subjected to multivariable logistic regression analyses. A variance inflation factor of below 5 and tolerance of above 0.2 indicated insignificant collinearity among independent variables. Correlation analysis of the non-normally distributed data was done by Spearman’s correlation coefficients. p < 0.05 was considered statistically significant.

The median ages of subjects infected with Omicron subvariant BA.2.76 and BA.5.1 were 50 (35, 60.75) and 38 (34, 50) years (p = 0.001), respectively. Gender ratio (female to male) was similar between two groups (p = 0.394). Due to the higher unvaccinated rates in BA.2.76 group (8.86% vs. 1.16%, p = 0.022) and proved efficacy of SARS-CoV-2 vaccination on symptomatic infection (20, 21), we first analyzed the impact of vaccination status on clinical features of patients infected with BA.2.76 or BA.5.1 and observed comparable COVID-19 severity and symptoms between unvaccinated and vaccinated participants (Supplementary Table S1). Of note, patients infected with BA.5.1 had higher incidence of clinical manifestations including fatigue (25.58% vs. 2.53%, p < 0.001), dyspnea (5.81% vs. 0%, p = 0.005), abdominal pain (3.49% vs. 0%, p = 0.049), headache or dizziness (12.79% vs. 4.43%, p = 0.017), nausea or vomiting (10.47% vs. 1.27%, p = 0.002), and higher viral load (CT of nucleocapsid protein (CT-N): 18.16 vs. 19.82, p = 0.001; CT of open reading frame (CT-O): 18.14 vs. 19.43, p = 0.010), but shorter period for SARS-CoV-2 clearance (10 vs. 11 days, p = 0.028) compared with patients with BA.2.76 infection. The differences in clinical features at onset may result from the different Omicron subvariants infection.

Additionally, the absolute value of lymphocytes (p < 0.001), and the levels of tumor necrosis factor (TNF)-α (p = 0.024), and IL-6 (p = 0.026) were higher in BA.5.1 group than BA.2.76 group, while the titers of anti-SARS-CoV-2 IgG and IgM titers was comparable between the two groups at baseline (Table 1). Other laboratory indexes with no significant difference were displayed in Supplementary Table S2.

The safety of Azvudine was evaluated in this current study. Adverse events had occurred in 28.1% (34/121) of the patients following Azvudine administration, such as nausea (13/121, 10.74%), diarrhea (7/121, 5.76%), vomiting (3/121, 2.47%), and headache (1/121, 0.83%). Notably, no serious adverse events (AEs) were observed for the group that received Azvudine (Table 2). Additionally, AE was found to be closely related to IL -1β levels before treatment, age, and gender (Supplementary Figure S1A). Patients with AEs, or AEs of vomit exhibited higher levels of IL-1β before treatment than those without (Supplementary Figures S1B,C). Older patients were more likely to have AEs of upset upset stomach (Supplementary Figure S1D). Nausea and vomit after Azvudine administration were more common in female patients, while liver injury was the opposite (Supplementary Figures S1E–G).

All patients exhibited favorable prognosis, who were discharged after treatment. Patients in Azvudine group had lower proportion of asymptomatic disease (25.62% vs. 46.34%, p = 0.001) and higher moderate severity (18.18% vs. 2.44%, p < 0.001) than those in control group (Supplementary Table S3). Azvudine drug still exhibited great benefits for virus clearance, the enhancement of anti-SARS-CoV-2 antibodies and immune cells, as well as the regulation of inflammatory factors. Shorter nucleic acid negativization (NAN) durations was observed in Azvudine rather than control group (10 vs. 11 days, p = 0.032) (Figure 1A). The levels of anti-SARS-CoV-2 IgG and IgM, white cell counts (WBC), lymphocyte counts, and platelets (PLT) were elevated to varying degrees, whose fold changes (FC) in Azvudine group were greater than that in control group (Figures 1B,C,G–I). Meanwhile, among patients receiving Azvudine, the FC of CRP, PCT and D-dimer markedly decrease compared with patients in control group (Figures 1D–F).

Given that patients infected with BA.2.76 or BA.5.1 exhibited varying periods for SARS-COV-2 clearance (Table 1), we conducted a detailed analysis of viral clearance times, as well as alterations in anti-SARS-CoV-2 antibodies, inflammatory factors, and immune cells, specifically comparing patients who received Azvudine treatment to those in the control group within both BA.2.76 and BA.5.1 cohorts (Supplementary Figures S2, S3). Although the differences were not statistically significant, patients treated with Azvudine had shorter NAN durations compared to those in the control group, both in the BA.2.76 cohort (11.07 vs.11.83 days) and the BA.5.1 cohort (10.14 vs. 10.91 days) (Supplementary Figures S2A–S3A). In BA.5.1-infected patients who received Azvudine, there were higher increases in anti-SARSCoV-2 lgG and lgM, WBC, lymphocyte counts, PLT, and greater reductions in CRP, PCT, and D-dimer levels compared to those in the control group (Supplementary Figures S3B–I). Patients infected with BA.2.76 showed similar trends except for WBC (Supplementary Figures S2B–I).

The correlation between the clinical and laboratory indexes and the duration of NAN was performed using Spearman’s correlation test (Figure 2A). Patients receiving Azvudine treatment (r = −0.14, p = 0.032), infected with BA.5.1 subvariant (r = −0.14, p = 0.027), without fever symptom (r = −0.13, p = 0.050) were negatively associated with SARS-CoV-2 clearance, and so is low lymphocyte counts (r = −0.27, p < 0.001) and anti-SARS-CoV-2 IgG titers (r = −0.20, p = 0.002) before treatment (Figures 1A, 2B–D). The shorter virus clearance time in the BA.5.1 group may be due to the Azvudine treatment they received.

Binary logistic regression analysis using model 1 showed that fever (OR = 1.942, 95% CI: 1.115–3.380; p = 0.019) was associated with >10 days NAN duration. Model 2 showed that low lymphocyte counts (OR = 0.607, 95% CI: 0.457–0.806; p = 0.001) and anti-SARS-CoV-2 IgG level (OR = 0.990, 95% CI: 0.990–0.999; p = 0.028) before treatment were independent risk factors for a prolonged NAN duration (Figures 2E–F).

In this study, 10.25% (25/244) individuals developed pneumonia. Patients with pneumonia exhibited a higher risk of dyspnea (12% vs. 0.91%, p = 0.008), fatigue (24% vs. 9.13%, p = 0.022), and a higher level of D-dimer (397 vs. 307 mg/L, p = 0.029) than patients without pneumonia (Figures 3A–C; Supplementary Table S3).

Then above three indicators were included in univariate logistic regression. Dyspnea (OR = 14.795, 95% CI: 2.345–93.361; p = 0.004) and fatigue (OR = 3.142, 95% CI: 1.126–9.771; p = 0.029) were associated with pneumonia. Risk factors with statistically significant differences in the univariate logistic regression were further analyzed in multivariate logistic regression, which further proved that dyspnea (OR = 10.176, 95% CI: 1.475–70.214; p = 0.019) was independent factors for the occurrence of pneumonia in patients with Omicron infection (Figures 3D–E).