The COVID-19 pandemic, caused by SARS-CoV-2, led to a significant number of individuals recovering from the infection. However, a substantial proportion of these individuals experience long-term effects, often referred to as Post-COVID Syndrome (PCS). Understanding the long-term humoral immune response and its correlation with PCS is crucial for developing effective strategies for managing and preventing this condition. Previous studies have shown antibody dynamics in the early stages post-infection, with seroconversion occurring 7-14 days after a positive qRT-PCR test and peaking around 30-35 days post-symptom onset. Long-term immunity has been reported for up to 6 months, but data beyond 12 months, particularly in the context of PCS, is limited. This study focused on the humoral immune response, specifically targeting anti-RBD/S1 and anti-N antibodies, due to the established correlation between RBD binding and protective potency. The study also investigated the impact of post-infection vaccination on PCS development and the potential use of cell-free DNA (cfDNA) as a biomarker for PCS. The primary aims were to examine the 12-month antibody titers in unvaccinated and vaccinated recovered subjects, identify parameters correlating with antibody decline, and assess the prevalence and symptoms of PCS at 12 months post-infection, exploring the potential of cfDNA as a predictive biomarker.

Existing literature extensively documents antibody dynamics in the early post-disease stages of COVID-19, showing seroconversion within 7-14 days and peaking at 30-35 days post-symptom onset. Studies have also reported long-term immunity up to 6 months. However, information about long-term immunity beyond 12 months in the clinical context of PCS remains scarce. The correlation between the RBD of the SARS-CoV-2 spike protein and protective potency has been established, making anti-RBD/S1 antibodies a suitable focus for studying humoral long-term immunity. The effect of vaccination on pre-existing PCS also requires further investigation. Prior research suggests a potential reduction in PCS risk through vaccination before infection, but its impact on post-recovery vaccination is unclear. The use of cfDNA as a biomarker in various infectious diseases has shown promise, prompting its exploration as a potential indicator of organ damage in the late recovery phase of COVID-19.

The Immunitor Study enrolled SARS-CoV-2-infected subjects (aged 18+) at least three weeks post-recovery. Participants provided serum and lithium-heparin plasma samples at six timepoints: study enrollment (t0), 2 weeks (t1), 1 month (t2), 3 months (t3), 6 months (t4), and 12 months (t5). Anti-SARS-CoV-2 anti-N and anti-RBD/S1 antibodies were measured using electrochemiluminescence immunoassays (ECLIA). HLA typing was performed at t0. At t0, 9 mL EDTA blood was collected for cfDNA isolation and quantification. Subjects were categorized into several cohorts: V1 (unvaccinated at t4), V1.1 (unvaccinated at t5), V1.2 (vaccinated between t4 and t5), a training cohort (t4 data from V1.2), and a validation cohort (V1.1 at t5). Antibody decline was defined as a >10% reduction from the previous value (trend 1), while stable or increasing titers constituted trend 2. A random forest algorithm was used to identify clinical predictors of antibody decline using the t4 training cohort, validated with the t5 validation cohort. Clinical data was collected via REDCap, and PCS symptoms were assessed at t4 and t5.

Forty-nine participants completed the 12-month follow-up. At t5, anti-RBD/S1 antibodies were detected in 48/49 and anti-N antibodies in 46/49 subjects. In the unvaccinated group (V1.1), anti-RBD/S1 antibodies remained largely stable, with only one subject showing a >50% decrease. However, anti-N antibody levels decreased significantly in 85% of subjects. The vaccinated group (V1.2) showed a mean 235-fold increase in anti-RBD/S1 antibodies post-vaccination. A random forest algorithm identified ‘other diseases’ and ‘vascular disease and other diseases’ as the most significant predictors of anti-RBD/S1 antibody decline, though the predictive power was limited. PCS prevalence was 38.6% at t5, with a notable increase in cognitive impairments. cfDNA levels did not significantly differ between subjects with and without PCS.

The stable anti-RBD/S1 antibody levels in the unvaccinated group suggest robust humoral immunity against SARS-CoV-2 beyond 1 year post-infection. The significant increase in antibody titers after vaccination in recovered individuals highlights the substantial booster effect of vaccination. The decline in anti-N antibodies underscores the differing dynamics of these antibody responses. The prediction model, although showing limited accuracy, suggests that pre-existing health conditions may play a role in antibody decline. The high prevalence of PCS at 12 months, particularly the increase in cognitive symptoms, emphasizes the significant long-term consequences of COVID-19. The lack of a significant difference in cfDNA levels between PCS and non-PCS groups suggests that cfDNA is not a reliable early biomarker for PCS.

This study demonstrates the persistence of anti-RBD/S1 antibodies in unvaccinated individuals for at least 12 months post-infection and the significant booster effect of post-recovery vaccination. The high prevalence of persistent PCS, particularly cognitive impairments, underscores the long-term impact of COVID-19. While the developed prediction model requires further refinement, this study highlights the need for comprehensive long-term follow-up studies to better understand the clinical implications of COVID-19 and develop effective strategies to mitigate long-term complications. Future research should focus on larger cohorts, incorporate more detailed clinical data, and explore other potential biomarkers for PCS.

The relatively small sample size of the unvaccinated cohort (due to high vaccination rates), reliance on self-reported symptoms, and the limited predictive power of the clinical prediction model are limitations. The study also lacked objective measures of cognitive function. Further research with larger, more diverse cohorts and validated cognitive assessment tools is needed to strengthen these findings and develop more precise prediction models.