The retrospective study was conducted at the Qingdao University Medical College Affiliated Yantai Yuhuangding Hospital, China, from December 15, 2022, to February 28, 2023. The inclusion criteria for the study were as follows: a) all patients who tested positive for SARS-CoV-2 infection via Real Time-Polymerase Chain Reaction (RT-PCR), b) CT imaging findings that met the criteria for viral pneumonia, and c) all the patients received azvudine treatment and underwent peripheral blood lymphocyte subset testing via flow cytometry. The exclusion criteria included: a) individuals under the age of 18, b) patients who received additional antiviral drugs, such as Paxlovid, alongside azvudine, c) those for whom flow cytometry measurement was not performed, and d) patients with incomplete data. At present, according to the Diagnosis and Treatment Program for Novel Coronavirus Pneumonia (ninth Edition) (20), the hospitalized COVID-19 patients were categorized into three groups based on the severity of their condition: moderate, severe, and critical. The present study was approved by the ethics committee of Yantai Yuhuangding Hospital (no. 2024521), and the requirement for informed consent was waived due to the retrospective study design.

The enrolled patients were administered azvudine treatment upon admission to the hospital and received additional treatments based on medical professionals’ discretion. These therapies included systemic corticosteroids, antibiotics, anticoagulants, immunoglobulin, supplemental oxygen including nasal catheter for oxygen/face mask oxygen inhalation, high-flow oxygen/noninvasive ventilation (HF/NIV), and invasive mechanical ventilation (IMV). Data derived from the electronic health records of COVID-19 included age, gender, clinical manifestations, medical history (e.g., comorbidities), imaging data, treatment regimens, clinical outcomes, and laboratory findings. The laboratory findings included lymphocyte subset parameters, interleukin-6 (IL-6), ferritin, C-reactive protein (CRP), procalcitonin (PCT), D-dimer, among others, and were conducted upon admission. Following the collection of blood specimens, azvudine administration was initiated. Lymphocyte subset analysis was conducted on patients utilizing antibodies and a cell analyzer supplied by Becton, Dickinson and Company.

The primary outcome was defined as all-cause mortality, along with a composite outcome of disease progression including symptom and CT finding aggravation, receiving more higher level of oxygen treatment, admission to the intensive care unit (ICU), and all-cause mortality. The secondary outcomes encompassed each of the individual disease progression. Patient outcomes were documented from admission until the occurrence of outcome events, discharge, or death, whichever occurred first. The discharge criteria included viral nucleic acid shedding and stable condition of any concurrent diseases.

All enrolled patients were followed for a period of 60 days from their admission, utilizing medical records and telephone consultations as data sources. This study primarily focused on the survival status throughout the 60-days follow-up period, with particular emphasis on accurately documenting the time of mortality.

Descriptive statistics were conducted to summarize all variables in the study. All continuous variables were presented as median (25th–75th percentile) and compared with the Mann–Whitney U test, while categorical variables were presented as numbers (%) by χ2 test or the Fisher’s exact test. Spearman rank correlation coefficient was employed to examine the associations between lymphocyte subsets and clinical indicators. A receiver operating characteristic (ROC) curve was plotted to determine the area under the curve (AUC), as well as the sensitivity and specificity of lymphocyte subset parameters. The association between the indicators and outcomes was estimated using Kaplan-Meier analysis and a multivariate logistic regression model. Statistical analyses were performed using SPSS 18.0 software (IBM Corp., Armonk, NY, USA) and GraphPad Prism 8.0 software (GraphPad Software Corp, San Diego, California, USA). A two-sided p < 0.05 was statistically significant.

Following established inclusion and exclusion criteria, data were ultimately collected and analyzed information from a cohort of 268 COVID-19 patients. Flow cytometry measurements were performed on all patients, followed by administration of azvudine treatment, and the patients were followed over a period of 60 days. The demographic, clinical characteristics, treatments regimens and outcomes of the participants were presented in Table 1 . The median age of the 268 enrolled patients was 72.00 (64.00–81.00) years, with a predominance of male patients (n = 174, 64.90%). Among the cohort, 50 patients reported a history of smoking history. The majority of participants had preexisting medical conditions, with hypertension (41.42%), diabetes mellitus (27.99%), coronary artery disease (19.78%), chronic pulmonary disease (13.43%), and cancer (7.46%) being the most prevalent. Chronic pulmonary diseases included asthma, chronic obstructive pulmonary disease, interstitial lung disease, and bronchiectasis. The proportion of enrolled patients classified as severe or critical was 42.91%. Notably, 207 (77.24%) received oxygen therapy. Systemic corticosteroids were administered to 247 patients (92.16%), while antibiotics were prescribed to 226 individuals (84.33%), and anticoagulants were utilized in 148 cases (55.22%). A minority of patients, specifically 17 individuals (6.34%), underwent immunoglobulin therapy.

The COVID-19 patients were categorized into two distinct groups based on a composite outcome of disease progression: namely the composite outcome group (n = 66, 24.63%) and the non-composite outcome group (n = 202, 75.37%) ( Table 1 ). Notable differences were observed between the two groups concerning age, duration of hospitalization, and arterial partial pressure of oxygen (PaO2)/fraction of inspired oxygen (FiO2) levels (all p < 0.05). Furthermore, the prevalence of hypertension, chronic pulmonary disease, and cancer among hospitalized patients was significantly higher in the composite outcome group compared to the non-composite outcome group (all p < 0.05). The occurrence of critical COVID-19 cases was also significantly higher in the composite outcome group compared to the non-composite outcome group (all p < 0.05). However, no statistically significant difference was found in the incidence of severe COVID-19 (p = 0.051). Although the p-value was not < 0.05, there was a trend that the severe type of COVID-19 was more common in composite outcome group. It is posited that an increase in the sample size could yield statistically significant results. Based on medical records, all patients received systemic steroid and antibiotics treatment in the composite outcome group. In addition, 49 (74.24%) patients received anticoagulants therapy and 17 (25.76%) patients received immunoglobulin therapy in the composite outcome group.

According to the final outcome, patients were classified into two groups - discharge (n = 231, 86.19%) or mortality (n = 37, 13.81%) ( Table 1 ). The median age and duration of hospitalization for the all-cause mortality group were significantly higher compared to those of the discharge group (p = 0.002 and p = 0.031 respectively). Additionally, the prevalence of chronic pulmonary disease was markedly higher in the all-cause mortality group compared to the discharge group (p = 0.003). The incidence of critical COVID-19 cases was significantly higher in all-cause mortality group relative to the discharge group (all p < 0.05). However, no statistically significant difference was observed in the prevalence of severe COVID-19 between the two groups (p = 0.055). Despite not reaching a significance level of p < 0.05, there was an observable trend suggesting a potential association between severe COVID-19 and patient mortality. Furthermore, the rates of interventions such as oxygen/face mask oxygen (NC/FM), high-flow oxygen/noninvasive ventilation (HF/NIV), and tracheal intubation (TI) were significantly higher in the all-cause mortality group compared to the discharge group (all p < 0.001). In addition, the administration of antibiotics, anticoagulants, and immunoglobulin treatment was more frequent in the all-cause mortality group (all p < 0.05).

The detailed characteristics of immune cells and lymphocyte subpopulations profile on admission of the COVID-19 patients were presented in Table 2 . The composite outcome group and all-cause mortality group demonstrated a statistically significant reduction in lymphocyte (L), monocyte (M), and platelet (PLT) counts. Conversely, there was a notable increase in neutrophil count (N) and neutrophil-to-lymphocyte ratio (NLR) when compared to the non-composite outcome group and the discharge group, respectively (all p < 0.05). The quantification of lymphocyte subsets was conducted using flow cytometry measurement ( Table 2 ). The counts of CD3⁺ T cell, CD4⁺ T cell, CD8⁺ T cell, B cell, and NK cell in both the composite outcome group ( Figure 1A ) and mortality group ( Figure 1B ) were significantly lower than the non-composite outcome group and the discharge group (all p < 0.001).

Inflammatory markers, along with the indicators of coagulation, cardiac function, and hepatorenal function were assessed in the enrolled patients ( Table 2 ). Notable statistical differences were observed between the non-composite outcome group and the composite outcome group for the following parameters: C-reactive protein (CRP), serum amyloid A (SAA), erythrocyte sedimentation rate (ESR), interleukin-6 (IL-6), D-dimer, creatine kinase-MB (CK-MB), cardiac troponin-I (cTn-I), lactate dehydrogenase (LDH), brain natriuretic peptide (BNP), albumin (ALB), alanine aminotransferase (ALT), blood urea nitrogen (BUN) and creatinine (Cr) (all p < 0.05). Additionally, a statistical difference was noted between the discharge and mortality groups for the following parameters: CRP, SAA, procalcitonin (PCT), ferritin, IL-6, D-dimer, CK-MB, cTn-I, LDH, BNP, aspartate aminotransferase (AST), ALB, BUN, and Cr (all p < 0.05).

The counts of CD3⁺ T cell, CD4⁺ T cell, CD8⁺ T cell, B cell and NK cell exhibited positive correlations with lymphocyte percentage (all p < 0.05) ( Figure 2 ). Conversely, these cell counts exhibited negative associations with neutrophil percentage, CRP, and ESR (all p < 0.05). Additionally, the levels of BNP were negatively correlated with the counts of CD3⁺ T cell, CD4⁺ T cell, CD8⁺ T cell, and NK cell (all p < 0.05), while no significant association was observed between B cell count and BNP levels. Furthermore, the counts of CD3⁺ T cell, CD4⁺ T cell, CD8⁺ T cell, B cell, and NK cell did not show significant associations with white blood cell (WBC) counts, platelet counts, albumin levels, creatinine levels, and PaO₂.

The predictive power of lymphocyte subsets for predicting the composite outcome and all-cause mortality in enrolled COVID-19 patients was evaluated using receiver operating curve (ROC)/area under the curve (AUC) through plotting sensitivity against specificity. As depicted in Figure 3A , the AUC derived from CD4⁺ T cell count demonstrated superior predictive value for the composite outcome compared to other lymphocyte subsets cells. The ROC curve for CD4⁺ T cell count, as presented in Figure 3A , yielded an AUC of 0.801 (95% CI = 0.731–0.871). Employing the maximum Youden index, an optimal cut-off value of 203.50 cells/μl was determined with a sensitivity of 91.70% and specificity of 60.70% ( Table 3 ). Similarly, as shown in Figure 3B , AUC derived from CD4⁺ T cell count exhibited the highest predictive value for all-cause mortality among different lymphocyte subsets. The ROC curve for CD4⁺ T cell count in assessing all-cause mortality is depicted in Figure 3B , yielding an AUC value of 0.765 (95% CI = 0.679–0.850). Utilizing the maximum Youden index, an optimal cut-off value of 156.00 cells/μl was identified with a sensitivity of 90.90% and specificity of 50.00% ( Table 3 ).

Based on the ROC curve analysis for all-cause death, the optimal cutoff value for CD4⁺ T cell count was determined as 156.00 cells/μl using Youden’s index. Subsequently, the enrolled patients were categorized into two distinct groups: Group A comprised individuals with CD4⁺ T cell count below 156.00 cells/μl, while Group B consisted of those with CD4⁺ T cell count equal to or exceeding 156.00 cells/μl. The enrolled patients were followed for a period of 60 days from the time of admission. The Kaplan-Meier curves revealed a significantly higher probability of mortality in patients with low CD4⁺ T cell count (< 156.00 cells/μl) compared to those with high CD4⁺ T cell count (≥ 156.00 cells/μl) (log-rank p < 0.001, HR = 8.242, 95% CI = 3.134–21.670) ( Figure 4 ).

A logistic regression analysis was conducted to identify lymphocyte subsets associated with composite outcome and mortality in COVID‐19 patients treated with azvudine ( Table 4 ). The univariable analyses revealed that low levels of CD3⁺ T cell, CD4⁺ T cell, CD8⁺ T cell, B cell, and NK cell were significantly associated with disease progression and mortality in the enrolled patients (all p < 0.001) ( Table 4 ). After adjusting for age, gender, comorbidities and severity at admission, the multivariate logistic regression analysis demonstrated that low levels of CD3⁺ cell (< 565.50 cells/μl), CD4⁺ cell (< 203.50 cells/μl), and NK cell (< 147.50 cells/μl) were significantly associated with disease progression in COVID-19 patients treated with azvudine (OR = 4.198, 95% CI = 1.155–15.262, p = 0.029; OR = 4.313, 95% CI = 1.645–11.307, p = 0.003; OR = 3.345, 95% CI = 1.436–7.791, p = 0.005; respectively) ( Table 4 ). Additionally, the multivariate logistic regression analysis revealed that low CD4⁺ T cell count (< 156.00 cells/μl) was associated with all-cause mortality in COVID-19 patients treated with azvudine (OR = 5.860, 95% CI = 1.727–19.878, p = 0.005).