A total of 259 patients were included in this study, including 110 patients with B.1 variant infection admitted to Changchun Infectious Disease Hospital from January to February 2021, 82 patients with BA.2 variant infection admitted to Department II of Jilin University Hospital from April to May 2022, and 67 patients with BA.5 variant infection admitted to Mehekou City Hospital from December 2022 to February 2023. The demographic data, clinical and laboratory parameters, and clinical classification were obtained through the hospital’s electronic medical record and analyzed by a group of skilled doctors.

In our study, clinical data included patient demographics (sex, age); “combined underlying disease” refers to pre-existing chronic conditions in patients, such as hypertension, diabetes, cardiovascular disease, pulmonary disease, renal disease, liver disease, neurological disease, and prior history of malignancy; signs and symptoms (fever, cough, chest tightness, weakness, palpitations, dyspnea, shortness of breath after activity, dizziness and headache, nausea and vomiting, nausea and diarrhea, asymptomatic), and duration of hospitalization.

Laboratory parameters included SARS-CoV-2 nucleic acid test results (ORF1ab gene and N gene Ct values) from the patient’s oropharyngeal swab and blood tests selected based on their established clinical relevance for COVID-19 severity assessment: (1) Renal function markers: creatinine (CREA, reference interval 41–73 10⁹/L): elevated levels correlate with acute kidney injury, a known complication of severe COVID-19 (Cheng et al., 2020). (2) Inflammatory and hematologic markers: white blood cells (WBC, 3.52–9.50 × 10⁹/L) and differential counts (NE% 40–75%, LY% 20–50%): lymphopenia and neutrophilia are hallmarks of SARS-CoV-2-induced cytokine storm (Terpos et al., 2020), and Platelet count (PLT, 125–350 × 10⁹/L): thrombocytopenia predicts poor outcomes in COVID-19 (Lippi et al., 2020). (3) Coagulation parameters: D-dimer (D-D, 0.00–0.50 mg/L): key predictor of thrombotic complications (Arachchillage and Laffan, 2020). APTT (28.0–42.0 s) and PT (11.0–15.0 s): coagulopathy indicators associated with disease severity (Levi et al., 2020). (4) Nutritional marker: Albumin (ALB, 40–55 × 10⁹/L): hypoalbuminemia reflects systemic inflammation and predicts mortality (Huang et al., 2020).

The disease is divided into four categories in clinical practice: mild, moderate, severe, and critical. Mild cases are those with mild clinical symptoms and no imaging signs of pneumonia; moderate cases are those with fever and respiratory symptoms and imaging signs of pneumonia; and severe cases are those with one of the following criteria: (a) dyspnea with respiratory rate ≥ 30/min; (b) pulse oximetry at rest ≤93%; (c) oxygenation index (arterial partial pressure of oxygen/inhalation oxygen fraction, PaO₂/FiO₂) ≤300 mmHg; (d) progressive clinical symptoms with significant progression of lesions >50% on lung imaging within 24–48 h; critical cases are those that meet one of the following criteria: (a) respiratory failure requiring mechanical ventilation; (b) shock; (c) combined with other organ failure requiring ICU monitoring and treatment. All diagnostic and clinical classifications of COVID-19 above are based on the “Diagnosis and Treatment Protocol for Novel Coronavirus Pneumonia (8th Trial Version)” published by the National Health Commission of the People’s Republic of China (Wei, 2020).

In this study, pharyngeal swabs from patients with COVID-19 were collected for routine COVID-19 diagnosis, and the inclusion criterion was a Ct value of <38 for nucleic acid testing in patients in 2021–2022, and <35 for nucleic acid testing in 2023. The negative criteria for nucleic acid testing in 2021–2022 were established as two consecutive nucleic acid tests for the ORF1ab gene and the N gene with a Ct value of ≥38 (fluorescence quantitative PCR method, two sampling intervals of at least 24 h); the negative criteria for nucleic acid testing in 2023 were established as two consecutive nucleic acid tests for the ORF1ab gene and the N gene with a Ct value of ≥35 (fluorescence quantitative PCR method, two sampling intervals of at least 24 h). When patients were admitted to the hospital, blood tests were conducted every 3 days for mild patients and as often as the dynamics of the disease permitted until the patients were discharged for severe patients. Four clinical classification were present in the cases included in this study, but there were not enough patients with the moderate or critical types to form separate groups. As a result, the text has referred to the combination of mild and moderate patients as “mild” and the combination of severe and critical patients as “severe.”

Laboratory parameters included ORF1ab gene and N gene Ct value testing for SARS-CoV-2 in patient oropharyngeal swabs and blood tests. Oropharyngeal swab sample collection and RT-qPCR assays were performed by trained medical personnel according to standardized procedures and traceability of sample results. “Guidelines for the Implementation of Regional 2019-nCoV Nucleic Acid Detection (2nd Edition)” and “Working Manual of Novel Coronavirus Nucleic Acid Detection for Medical Institutions (second trial Edition)” were strictly followed during the sample collection procedure. RNA extraction was performed using a nucleic acid extraction (NAE) system and accompanying NAE or purification reagents, and the operating procedures strictly followed the manufacturer’s protocol. Nucleic acid amplification was performed on a RT-qPCR detection system. The PCR reaction system configuration, reaction parameters, and program settings were set according to the kit’s instructions, respectively. Each run was subjected to quality control, which included three weakly positive and one weakly negative controls to detect false positive and false negative results. At the three hospitals, the blood testing equipments were put through stringent quality control testing. The Jilin Clinical Laboratory Center accredited all three of the study’s participating labs for inter-laboratory quality evaluation and proficiency. The Health Commission of Jilin Province provided standardized training to all of the doctors, technicians, and nurses involved in this study. For common biochemical analytes and blood cell analysis in Chinese adults, reference interval standards have been issued by the National Health Commission of the People’s Republic of China. Reference intervals that met the requirements were used by the three hospitals. There was no influence of different instruments on the test results.

All statistical analyses were performed using IBM SPSS Statistics version 22 (IBM Corporation, Armonk, NY, United States). Data distribution normality was assessed using the Shapiro–Wilk test. Continuous variables including Ct values, nucleic acid conversion time, and hematological parameters were analyzed as follows: for normally distributed data: group comparisons were conducted using unpaired Student’s t-tests, results expressed as mean ± standard deviation. For non-normally distributed data: Mann–Whitney U tests were employed, results presented as median (interquartile range, IQR). Categorical variables (demographic characteristics and clinical manifestations) were expressed as frequencies (percentages) and compared using: χ² tests for expected cell counts >5; Fisher’s exact tests for expected cell counts ≤5. Correlation analyses were performed using: Pearson’s correlation for normally distributed continuous variables; Spearman’s rank correlation for non-parametric data; binary logistic regression for dichotomous outcomes. All statistical tests were two-tailed, with p < 0.05 considered statistically significant. Graphical representations were generated using GraphPad Prism 8 (GraphPad Software, San Diego, CA) and OriginPro 2022 (OriginLab Corporation, Northampton, MA).

The data set was entirely anonymous and did not include any personally identifiable health information that might be used to violate the subjects’ rights or interests. The study was carried out in accordance with the approved guidelines and was approved by the ethics committees of the hospitals: Ethics Committee of Changchun Infectious Diseases Hospital (No. 2020-001) and First Hospital of Jilin University (No. K2022028). The Ethics Committee of the First Hospital of Jilin University (No. K2022028) and the Ethics Committee of Changchun Infectious Diseases Hospital (No. 2020-001) waived written informed consent for this study for it was a retrospective study.

Table 1 presents the demographic and clinical characteristics across the three SARS-CoV-2 variants. Key findings include: (1) Demographic distribution: No significant sex differences were observed among variants (p > 0.05). Significant age variations emerged: B.1 patients were younger (median: 56 years) than BA.2 (65 years) and BA.5 (72 years) cohorts (p < 0.05). (2) Comorbidity profiles: The prevalence of underlying diseases differed substantially (p < 0.05): B.1: 46.36% vs. BA.2: 67.07% vs. BA.5: 76.12%. (3) Disease severity: BA.5 variant demonstrated the highest proportion of severe cases (50.75%), contrasting with BA.2 (0%) and B.1 (8.18%) (p < 0.05). (4) Vaccination status: All B.1 variant cases were unvaccinated (100%). Comparable vaccination rates between BA.2 (45.12%) and BA.5 (34.33%) variants (p > 0.05).

Our analysis revealed distinct patterns in Ct values (representing viral load) among the three variants (Table 2 and Figure 1): (1) Inter-variant comparisons: B.1 variant demonstrated significantly lower median Ct values (higher viral loads) for both ORF1ab (median: 31.37) and N genes (median: 30.49) compared to BA.2 (ORF1ab: 33.00; N: 32.00; both p < 0.05). The N gene consistently showed lower Ct values than ORF1ab across all variants. (2) Clinical severity association: B.1 severe cases exhibited significantly lower viral loads than mild cases (ORF1ab ΔCt = 2.58; N gene ΔCt = 2.22, both p < 0.05). (3) Omicron subvariants (BA.2 vs. BA.5): BA.2 showed higher Ct values (lower viral loads) than BA.5 overall (ORF1ab: 33.00 vs. 32.38; N: 32.00 vs. 31.54; both p < 0.05), this pattern persisted in vaccinated subgroups (ORF1ab: 32.00 vs. 30.96; N: 32.00 vs. 30.23, both p < 0.05). Gender differences were observed in BA.2 (Male vs. Female: ORF1ab 33.00 vs. 32.00; N 33.00 vs. 31.00, both p < 0.05). (4) Vaccination effects: Paradoxically, BA.5 vaccinated patients showed lower Ct values than unvaccinated (ORF1ab: 30.96 vs. 34.66; N: 30.23 vs. 33.34, both p < 0.05). (5) Longitudinal patterns: All variants showed progressive Ct value increases (viral load decline) over time (Figure 1).

Results of viral clearance kinetics analysis revealed significant differences among variants (Figure 1). The median time to negative nucleic acid conversion was longest in B.1 variant patients (18 days), followed by BA.5 (15 days) and BA.2 variants (14 days), with B.1 demonstrating significantly prolonged clearance compared to both Omicron subvariants in all stratified analyses (p < 0.01). Multivariate analysis showed that: for B.1 and BA.5 variants: no significant associations were observed between nucleic acid conversion time and sex, age, disease severity, comorbidities, or vaccination status (all p > 0.05). For BA.2 variant: advanced age correlated with prolonged viral clearance (p < 0.05). Unvaccinated status was associated with longer conversion time compared to vaccinated individuals (p < 0.05) (Figure 2). Notably, no significant inter-group differences were observed between BA.2 and BA.5 variants in other demographic or clinical subgroups (all p > 0.05). These findings suggest variant-specific patterns of viral persistence, with BA.2 showing unique age- and vaccine-dependent clearance dynamics.

Clinical symptom profiles differed significantly among the three variants (Table 3). While fever and cough represented the most prevalent symptoms across all variants, their clinical manifestations exhibited distinct patterns: (1) Asymptomatic presentation: The B.1 variant demonstrated the highest proportion of asymptomatic cases (n = 91, 82.73%), significantly exceeding BA.2 (n = 14, 17.07%) and BA.5 (n = 3, 4.48%) variants (p < 0.001). (2) Symptom spectrum: BA.2 and BA.5 variants were associated with broader symptom profiles, including: cardiorespiratory symptoms: chest distress (B.1: 1.82% vs. BA.2: 13.41% and BA.5: 35.82%, p < 0.05) and dyspnea (B.1: 0.91% vs. BA.2: 13.41% and BA.5: 16.42%, p < 0.05); Systemic manifestations: weakness (B.1:1.82% vs. BA.2: 20.73% and BA.5: 31.34%, p < 0.05) and myalgia (B.1: 0.91% vs. BA.2: 13.41% and BA.5: 16.42%, p < 0.05); Neurological symptoms: headache (B.1: 0 vs. BA.2: 12.20% and BA.5: 19.40%, p < 0.05). (3) Variant-specific features: BA.5 variant patients showed significantly higher rates of: shortness of breath after activity (BA.5: 26.87% vs. BA.2: 2.44%, p < 0.05); gastrointestinal symptoms (nausea/vomiting): (BA.5: 16.42% vs. BA.2: 2.44%, p < 0.05). These findings suggest evolutionary changes in viral tropism and pathogenicity across variants, with Omicron sublineages (particularly BA.5) exhibiting enhanced respiratory and systemic symptomatology.

To assess risk factors associated with disease severity and clinical manifestations, we performed multivariate logistic regression analyses incorporating sex, age, underlying comorbidities, and vaccination status across all three variants. Key findings demonstrated: (1) Disease severity predictors (Figure 3A): advanced age (≥60 years) emerged as a significant independent risk factor for severe disease classification (adjusted OR = 5.92, 95% CI: 2.02–17.38; p < 0.001), after controlling for sex, comorbidities and vaccination status. (2) Symptomatic presentation predictors (Figure 3B): the presence of underlying comorbidities showed the strongest association with symptomatic disease (adjusted OR = 6.61, 95% CI: 3.15–13.86; p < 0.01), independent of age, sex or vaccination status. Notably, vaccination status did not demonstrate statistically significant associations with either outcome measure in our regression models (all p > 0.05). These findings highlight the differential impacts of demographic versus clinical factors on COVID-19 outcomes during the study period.

Patients infected with the BA.2 and BA.5 variants showed significantly higher serum creatinine (CREA), white blood cell count (WBC), and neutrophil percentage (NE%) compared to those infected with the B.1 variant (p < 0.01), while albumin (ALB) and lymphocyte percentage (LY%) were significantly lower than in B.1 variant patients.

Among patients infected with the B.1 and BA.5 variants, as disease severity increased, LY% decreased significantly, whereas CREA, NE%, activated partial thromboplastin time (APTT), prothrombin time (PT), and D-dimer (D-D) increased significantly (Table 4). However, in patients infected with the B.1 variant, blood test parameters such as CREA, WBC, APTT, PT, NE%, platelet count (PLT), and LY% fluctuated within the reference ranges throughout the disease course. In contrast, in patients infected with the Omicron BA.2 and BA.5 variants, most measurements of CREA, ALB, D-D, and LY% were either above or below the reference ranges (Figure 4).

In this study, when patients with all three variants were combined for analysis, Pearson correlation analysis revealed that the number of days since infection was positively correlated with ALB, NE%, and PT, and negatively correlated with LY% (p < 0.01) (Table 5). Additionally, correlation analysis of patient sex, age, underlying diseases, vaccination status, clinical classification, and blood test parameters showed that age, underlying comorbidities, and COVID-19 clinical classification were negatively correlated with ALB and LY%, but positively correlated with WBC, NE%, and D-D. Furthermore, age was positively correlated with COVID-19 clinical classification and underlying comorbidities. The correlations between other indicators are shown in Figure 5.

The demographic and clinical profiles of patients varied significantly between variants, likely influenced by evolving public health policies. During the BA.5 wave (December 2022–February 2023), quarantine measures were lifted, leading to broader community transmission. This shift may explain the higher proportion of elderly and comorbid patients hospitalized with BA.5, consistent with reports that Omicron BA.5 causes milder disease in younger, healthier individuals but poses greater risks to older populations with underlying conditions (Yang et al., 2022; Li et al., 2020). Logistic regression confirmed that age >60 years was strongly associated with severe disease (OR: 5.919), aligning with global trends (Qassim et al., 2022).

Viral load dynamics, assessed via ORF1ab and N gene Ct values, revealed three key observations: (1) Temporal trends: Ct values plateaued early (Day 2 post-infection) before gradually rising, with BA.2 and BA.5 showing higher variability during recovery, possibly due to sampling limitations (Zhou et al., 2022). (2) Gene-specific differences: The N gene had lower Ct values than ORF1ab in B.1 infections, supporting its higher sensitivity in diagnostics (Zhuang et al., 2021). (3) Variant-specific viral load: While BA.5 and B.1 had comparable median Ct values, BA.2 exhibited significantly lower viral loads and shorter nucleic acid clearance times. Notably, BA.5 demonstrated enhanced replicative fitness over BA.2 (Ong et al., 2023), though vaccination status may confound these findings (see limitations).

The evaluation of factors influencing the negative conversion cycle among the three variants revealed that significant differences in the time to negative conversion were only observed in patients aged ≥60 years and unvaccinated status. In contrast, the presence of underlying diseases and disease severity showed no significant impact on the negative conversion cycle. This observation may be attributed to the following factors: (1) Variant-specific pathological mechanisms: Emerging evidence suggests that Omicron variants (BA.2/BA.5) may exhibit shorter tissue persistence compared to the ancestral strain, potentially reducing clearance differences associated with disease severity (Yang et al., 2023). (2) Study population characteristics: Our hospitalized cohort may represent a narrower spectrum of disease severity than the general COVID-19 population. (3) Treatment homogeneity: The standardized antiviral treatment protocol implemented in our hospital may have minimized clearance variations related to disease severity. (4) Sample size limitations: Subgroup analyses were powered at 80% to detect differences of ≥3 days. (5) Biological plausibility: Recent studies have also indicated that while vaccination accelerates viral clearance, disease severity may not significantly prolong viral shedding time in hospitalized patients during the Omicron era (Yang et al., 2023).