Clinical data and specimens were obtained from unvaccinated patients COVID-19 who were hospitalized or followed at Chosun University Hospital in Gwangju Metropolitan, South Korea. The individuals recruited for this study were patients with SARS-CoV-2 infection between February 2020 and February 2021, prior to the occurrence of the SARS-CoV-2 variant(s). For diagnosis, a molecular method was used, namely, our in-house-designed reverse transcription polymerase chain reaction (RT-PCR) targeting the N gene, and a commercial kit (Kogene Biotech Seoul, South Korea) targeting E and RdRp genes, according to the manufacturer’s protocol. COVID-19 was diagnosed when more than two genes were detected at a C_t value <38 or when the SARS-CoV-2 culture showed a positive result. For patients with a body temperature of ≥37.5°C, the time to fever clearance was defined as the period from the initial fever onset until the body temperature decreased to ≤37.3°C and remained below this temperature for at least 48 h without the use of an antipyretic. COVID-19 patients were divided into four groups based on disease severity: asymptomatic patients (n = 24), those who had no symptoms throughout the course of the infection; patients with mild-to-moderate illness (n = 36), those who were symptomatic but did not receive supplemental oxygen or supplemental oxygen via a nasal prong; severely ill patients (n = 14), those who required high-flow oxygen therapy; and critical/fatal cases (n = 23), those who required mechanical ventilation or died. Chest radiography was performed on the day of hospitalization. X-ray scores were obtained by dividing each lung into upper, middle, and lower zones and scoring each zone from 0 to 4 points based on the degree of infiltration. The scores for each lung (with a total of six zones) were summed to yield a total score of 0–24 (8). The study protocol was approved by the Institutional Review Board of Chosun University Hospital (approval no. CHOSUN 2020-11-007-003). Written informed consent was obtained from all the participants or their legal guardians. All methods were performed in accordance with relevant guidelines and regulations.

The IgG antibodies of patients with COVID-19 were detected using an indirect enzyme-linked immunosorbent assay (ELISA) with the SARS-CoV-2 spike S1 domain recombinant protein as the antigen. The plant-derived S1 recombinant protein was donated by BioApp (Pohang, South Korea). After adding 0.2 μg of S1 recombinant protein to each well of the ELISA plate overnight at 4°C, the wells were washed with phosphate-buffered saline containing 0.05% Tween 20 (PBS-T) solution, followed by the addition of 5% skim milk in PBS-T for blocking. The patient’s serum in 1:100 dilution was added three times to each well for reaction with the antigen protein at 37°C for 2 h. After washing, the diluted horseradish peroxidase-conjugated goat anti-human IgG (Thermo Fisher Scientific), as a secondary antibody, was incubated at 37°C for 1 h. After washing, 50 μL of 3,3′-, 5,5′-tetramethylbenzidine substrate solution was added to each well and incubated at 20–25°C for 30 min. Subsequently, 25 μL of 1 N H₂SO₄ was added to stop the reaction. Absorbance at 450 nm (A₄₅₀) was then measured. An A₄₅₀ value of mean + 3 standard deviations or more was established as the positive cutoff value using ELISA on clinical samples from 15 patients who underwent health screening. For each experiment, ELISA was performed by adding a positive control, negative control, and internal control with an A₄₅₀ value of 1.

To perform an indirect immunofluorescence assay (IFA), SARS-CoV-2 samples obtained from the Korea Centers for Disease Control and Prevention were used to infect Vero E6 cells. To prepare the SARS-CoV-2 antigen slide, cells infected for 3 days were cultured on Teflon-coated well slides overnight at 37°C in a 5% CO₂ environment and fixed with 80% acetone the next day. The patient’s serum was diluted using a twofold serial dilution from 1:16 and then reacted with SARS-CoV-2 antigens in a moist chamber for 30 min at 37°C. After washing, slides were incubated with secondary antibodies at a 1:400 dilution (fluorescein isothiocyanate-conjugated anti-human IgG; MP Biomedicals, OH, USA). The slides were examined under a fluorescence microscope (Olympus IX73, magnification: 400×) after dispensing the mounting solution (Vector Laboratories). An IgG antibody titer of ≥1:32 was established as the cut-off value using IFA on clinical samples from 15 individuals who underwent health screening.

For detection of SARS-CoV-2 neutralizing antibodies, 100 μL of fourfold serially diluted test serum was mixed with 100 μL of 100 plaque-forming units of the SARS-CoV-2 isolate BetaCoV/South Korea/KUMC01/2020 and incubated at 4°C for 1 h. The virus:serum mixture (100 μL) was added to Vero E6 cells (Korean Cell Line Bank, KCLB no. 21587), and adsorption was performed at 37°C in an incubator with 5% CO₂ for 1 h, after which a 1% methyl cellulose overlay prepared in cell culture maintenance medium (Dulbecco’s modified Eagle’s medium, 5% heat-inactivated fetal bovine serum) was applied. At 5 days post-infection, the inverse of the highest dilution of serum providing 50% (PRNT₅₀) viral plaque reduction relative to the virus-only infection was reported as the titer. A 1:10 dilution was considered the lowest possible significant titer. All cell culture infection experiments were performed in a biosafety level 3 laboratory at the Health and Environment Research Institute of Gwangju City and Seoul National University.

To quantify the neutralisation potency of humoral immune responses against SARS-CoV-2, we calculated the neutralization potency index (PRNT₅₀/IFA IgG) for each patient.

Statistical analyses were performed using GraphPad Prism 8.0.1 (GraphPad Software, San Diego, CA, USA) and SPSS Statistics 26 (IBM Corp., Armonk, NY, USA). Data on 50% viral plaque reduction relative to virus-only infection (PRNT₅₀) were analyzed at 50% inhibition concentration (IC₅₀) using GraphPad Prism 8.0.1. A nonparametric multivariate analysis of variance (ANOVA) was performed on the indicated figures. Statistical significance was defined as p<0.05. All correlations were analyzed using two-tailed Spearman’s tests; the r and p values are indicated in the figures. The Kaplan–Meier method was used to estimate seroconversion time using SPSS Statistics 26.

Antibody responses in 97 RT-PCR-confirmed individuals with COVID-19 were evaluated at sequential time points. Among the patients, 55.67% and 44.33% were men and women, respectively, with a mean age of 57 (range, 22–92) and 62.6 years (range, 20–93) years, respectively ( Figure 1A ). Patients with critical illness were significantly older (mean age: 75.3, p < 0.0001; Figure 1B ) than those with asymptomatic or mild-to-moderate illness and had a longer duration of fever (mean number of days: 14.61, p = 0.0003) than those with asymptomatic-to-severe illness. Baseline characteristics of the disease severity groups are summarized in Table 1 and Figure 1 .

A total of 314 serum samples from 97 patients were tested using the PRNT, and the highest serum dilution that reduced plaque numbers by 50% (PRNT₅₀) was determined using the detection limit of the 1:10 antibody titer. The 35 serum samples from healthy subjects contained neutralizing antibodies at a titer <1:10 ( Figure 1F ). Of the 314 serum samples collected from individuals 0 to 430 days after symptom onset, 85.35% tested positive for the PRNT₅₀ antibody test.

The neutralizing antibody titer was 111 ± 35.34 (mean ± SD; positivity rate: 58.14%) at 1 week after symptom onset and increased to 500 ± 105.5 (positivity rate: 87.1%) at 2 weeks ( Figures 1F, G , Supplementary Table 1 ). The neutralization peak occurred at 15–42 days after symptom onset (PRNT₅₀ antibody titer: 1120 ± 161.6), where 97.8% of serum samples tested positive in the PRNT₅₀ antibody titer test ( Figures 1F, G , Supplementary Table 1 ).

After 43 days, the number of serum samples with high PRNT₅₀ titers decreased significantly. Of the 41 serum samples, 32 (78.05%) collected within 241–430 days after symptom onset remained positive for the PRNT₅₀ antibody titer at 62.6 ± 18.36 (mean ± SD; Figures 1F, G , Supplementary Table 1 ). Furthermore, men had higher concentrations of neutralizing antibodies in serum than females did (1435 ± 271.6 versus 753.4 ± 131.5, p = 0.0343) at 15–42 days after symptom onset ( Figure 1D , Supplementary Table 2 ). The neutralizing antibody level in patients aged ≥65 years was higher than that in patients aged <65 years, 15–120 days after symptom onset ( Figure 1E ).

The neutralizing antibody response depending on disease severity at different time points after symptom onset was as follows: the PRNT₅₀ antibody titers in the critical/fatal cases were significantly higher than those in the mild-to-moderate illness and asymptomatic groups at 15–42 days after symptom onset ( Figure 1H , Supplementary Table 4 ). At 1-year follow-up, the neutralizing antibody levels in the serum of asymptomatic patients were significantly lower than those of symptomatic patients (p = 0.0007; Figure 1I , Supplementary Table 3 ), particularly in critical/fatal cases (p = 0.0016; Figure 1H , Supplementary Table 4 ).

The neutralizing antibody response depending on antiviral drug treatment at different time points after symptom onset was as follow: the PRNT₅₀ antibody titers in patients treated with lopinavir/ritonavir or remdesivir were significantly higher than those in patients with non-antiviral treatment 15–42 days after symptom onset (p < 0.001; Supplementary Table 6 ). At 1-year follow-up, the neutralizing antibody levels were almost the same in the antiviral treatment and non-treatment groups.

The IgG response against SARS-CoV-2 was measured using IFA and ELISA at multiple time points ( Supplementary Figures 1A, B ). In total, 249 serum samples from 89 patients were tested via IFA using a SARS-CoV-2 antigen slide with a detection limit of a 1:32 antibody titer. The amount of IgG antibodies based on the IFA increased with time after symptom onset, peaking 15–42 days after symptom onset (IFA IgG titer: 864 ± 103.3; Supplementary Figure 1A , Supplementary Table 1 ). The IgG antibodies detected by IFA appeared to persist within the analyzed timeframe for up to 1 year. The IFA IgG antibody positivity rates were 41.03% at 1 week, 100% at 4 weeks, and 61.11% at 1 year after symptom onset. A total of 288 serum samples from 95 patients were tested through ELISA for the S1 antigen with an antibody titer detection limit corresponding to A₄₅₀ of 0.3. The amount of anti-S1 ELSIA IgG increased 4 months after symptom onset and showed a tendency to decrease gradually thereafter ( Supplementary Figure 1B , Supplementary Table 1 ). In contrast, the positivity rate of Anti-S1 ELISA IgG was 90% even at 1 year, unlike the IFA IgG result.

The relationships among all individual antibodies measured at the neutralizing antibody peak (15–42 days after symptom onset) and IFA IgG levels correlated with neutralization (r = 0.67, p < 0.0001; Figure 2A ). Anti-S1 IgG ELISA also showed a strong correlation with neutralization (r = 0.76, p < 0.0001; Figure 2B ). Anti-S1 ELISA IgG levels positively correlated with IFA IgG levels (r = 0.72, p < 0.0001; Figure 2C ). The humoral antibody response and neutralizing antibody titers 15–42 days after symptom onset were higher in the critical/fatal cases than in the mild-to-moderate-illness group ( Figures 2D–F ).

We analyzed the potential correlation between antibody levels at 15–42 days after symptom onset and several factors, including fever duration, the viral clearance period, maximum viral load, age, an X-ray score, the C-reactive protein (CRP) level (mg/dL), and the lymphocyte count ( Figures 2G–M ). Neutralizing antibody levels were significantly and positively correlated with fever duration, age, the X-ray score at the time of hospitalization, and the CRP level (mg/dL) within 1 week after symptom onset ( Figures 2G–J ) but not with the viral clearance period and maximum viral load in respiratory secretions during the initial infection phase ( Figures 2L, M ). Neutralizing antibody levels were negatively correlated with the lymphocyte count at 1 week after symptom onset ( Figure 2K ). The IgG antibody detected by the IFA was correlated with fever duration, maximum viral load, and age but not with the viral clearance period ( Supplementary Figures 2A–D ). The anti-S1 ELISA IgG antibody was correlated with fever duration but not with the viral clearance period, maximum viral load, and age ( Supplementary Figures 2E–H ).

Serum samples obtained from 60 patients who achieved seroconversion in terms of neutralizing or other humoral response antibodies were analyzed 30 days after symptom onset ( Figures 3B–D ). The kinetics of neutralizing antibodies at 1-year follow-up in asymptomatic and symptomatic patients are presented in Figure 3A . The number of neutralizing antibodies in asymptomatic patients decreased remarkably within one year of symptom onset. The results showed that the time to the emergence of the peak of the neutralizing antibody against SARS-CoV-2 was shorter in asymptomatic patients than in symptomatic patients (p = 0.0006; Figure 3A ). The median number of days to the seroconversion of neutralizing antibodies in asymptomatic patients was 10 days after symptom onset, which was shorter than that in symptomatic patients (15 days, p = 0.002; Figure 3B ). The median number of days to IgG IFA and anti-S1 ELISA IgG seroconversion in asymptomatic patients was 11 days after symptom onset, which was also earlier than in symptomatic patients (15 days, p = 0.004 and 16 days, p = 0.0004, respectively; Figures 3C, D ). The median number of days to neutralizing antibody seroconversion was lower in the asymptomatic group than in the mild-to-moderate illness (16 days, p = 0.003) and critical/fatal illness groups (15 days, p = 0.008; Figure S3A ). The median number of days for IgG IFA and anti-S1 ELISA IgG antibody seroconversion was similar to that for neutralizing antibody seroconversion ( Supplementary Figures 3B, C ).

This index was significantly lower in symptomatic than in asymptomatic patients at 1-year follow-up and at the neutralizing antibody peak ( Figures 3E, F ). Furthermore, the neutralization potency index was significantly lower in patients with mild-to-moderate illness and severely ill patients than in asymptomatic patients at 0–430 days after symptom onset ( Figure 3G ). In contrast, the neutralization potency index was higher in critical patients than in those with mild-to-moderate illnesses ( Figure 3G ). This index was significantly lower in severely ill patients at peak neutralizing antibody levels than in asymptomatic patients ( Figure 3H ).