Abstract
The factors involved in the persistence of antibodies to SARS-CoV-2 are unknown. We evaluated the antibody response to SARS-CoV-2 in personnel from 10 healthcare facilities and its association with individuals’ characteristics and COVID-19 symptoms in an observational study. We enrolled 4735 subjects (corresponding to 80% of all personnel) for three time points over a period of 8-10 months. For each participant, we determined the rate of antibody increase or decrease over time in relation to 93 features analyzed in univariate and multivariate analyses through a machine learning approach. In individuals positive for IgG (≥ 12 AU/mL) at the beginning of the study, we found an increase [p= 0.0002] in antibody response in symptomatic subjects, particularly with anosmia/dysgeusia (OR 2.75, 95% CI 1.753 – 4.301), in a multivariate logistic regression analysis. This may be linked to the lingering of SARS-CoV-2 in the olfactory bulb.
Introduction
Early in the COVID-19 pandemic, a major open question was the durability of natural immunity. While some studies indicated that antibodies faded quickly, particularly in asymptomatic individuals, the exact factors driving the persistence, decline, or increase of the SARS-CoV-2 antibody response remained largely unknown. This study investigates how long antibodies last and seeks to identify specific demographic, occupational, and symptom-based features that correlate with sustained or increasing antibody levels over time.
Methods
Study Design and Participants
This observational study tracked 4,735 healthcare workers across 10 facilities in Northern Italy. Blood samples were collected at three time points over an 8-10 month period to measure anti-S1/S2 IgG antibodies. The dataset captured 93 features per participant, including demographics, 72 categorical symptoms, and comorbidities.
Analytical Approach
Instead of absolute antibody levels, the researchers focused on the rate of change (increase or decrease) over time. To analyze the high-dimensional feature set, they employed:
- Statistical Modeling: Univariate and multivariate logistic regression to identify symptom correlations.
- Machine Learning: A Bagging classifier composed of 7 logistic regressions, optimized via Bayesian hyperparameter tuning. Given the high kurtosis of the antibody rate-of-change data, the model was trained to classify subjects in the extreme tails of the distribution (the <10th and >90th percentiles).
Results
Long-term Persistence of Antibodies
The study demonstrates that natural immunity is highly durable, persisting for at least 8 to 10 months. This challenges early concerns that immunity might vanish within a few months of infection.
Symptom-Driven Antibody Trajectories
A key finding was that symptom severity dictates the antibody trajectory. While asymptomatic individuals typically saw their antibody levels decrease, symptomatic patients experienced an increase in antibody rates between the first and second testing phases.
Specifically, anosmia/dysgeusia (loss of smell and taste) was the strongest predictor of a sustained or increasing antibody response:
- Multivariate Logistic Regression: Anosmia/dysgeusia was associated with an increased antibody rate with an Odds Ratio (OR) of 2.75 (95% CI 1.75–4.30).
- Extreme Increases: Among subjects in the top 10% for antibody increase rates, 69% had experienced loss of smell or taste.
- Chest Pain: This symptom also showed a strong correlation (OR 2.32), though 54% of these patients also reported anosmia/dysgeusia.
Demographic Trends
- Sex: Females sustained antibody responses slightly better than males ($p = 0.01$).
- Occupation: Non-medical healthcare workers showed higher antibody rates, possibly reflecting different viral exposure profiles compared to medical staff.
Model Performance
The Bagging classifier achieved a test accuracy of 72.00% (ROC AUC = 72.12, Recall = 81.08) in predicting extreme antibody rate changes based on the input features.
Discussion
The authors hypothesize that the strong correlation between anosmia and increasing antibody levels may be due to a viral reservoir. The virus may persist in the olfactory bulb long after clearing the respiratory tract, causing ongoing inflammation and antigen release that acts as a natural "booster" for the immune system.
Limitations and Future Work
The study is limited by its focus on anti-S1/S2 IgG antibodies, excluding other subtypes like IgA or IgM. Additionally, the hypothesis of a viral reservoir in the olfactory bulb requires direct tissue testing for confirmation. Future research should combine antibody tracking with direct testing for SARS-CoV-2 persistence in localized tissues.
These findings suggest that individuals who recovered from symptomatic COVID-19 might only require a single vaccine dose to achieve robust boosting, even months after their initial infection.
Summary updated