Background In the first weeks of the COVID-19 epidemic in Belgium, a repetitive national serum collection was set up to monitor age-related exposure through emerging SARS-CoV-2 antibodies. First objective was to assess the baseline seroprevalence and sero-incidence using serial survey data that covered the start of a national lock-down period installed soon after the epidemic was recognized.
Methods The prospective serial cross-sectional nationwide seroprevalence study, stratified by age, sex and region, contains 3000-4000 samples in each of the five collection periods (April till June 2020). In residual sera taken outside hospitals and collected by diagnostic laboratories, IgG antibodies against S1 proteins of SARS-CoV-2 were measured with a semi-quantitative commercial ELISA. Seropositivity (cumulative, by age category and sex) and sero-incidence over 3-4 weeks’ periods were estimated for the Belgian population.
Findings The weighted overall seroprevalence initially increased from 2,9% (95% CI 2,3 to 3,6) to 6,0% (95% CI 5,1 to 7,1), reflected in a sero-incidence estimate of 3,1% (95% CI 1,9 to 4,3) between 1st and 2nd collection period. Thereafter, seroprevalence stabilized and decreased from 3rd to 5th period from 6,9% (95% CI 5,9 to 8,0) to 4,5% (95% CI 3,7 to 5,4).
Interpretation During the start of epidemic mitigation by lockdown, a small but increasing fraction of the Belgian population showed serologically detectable signs of exposure to SARS-CoV-2. The decrease observed afterwards most probably reflects rapid waning of antibodies after infection.
Goal of the study
Collecting residual serum samples to detect antibodies against COVID19, see how the proportion of susceptible people in the population evolves over time and how this is distributed over age groups (and gender and region). In other words, to monitor how many people have been infected over time (cumulatively), and whether the evolution differs across age groups (and gender and region). This is very important information, both for the policy to control the current outbreak and for the prevention / treatment of a possible 2nd or 3rd wave.
2 laboratories that mainly serve outpatient practices, AML (Algemeen Medisch Laboratorium) and Laboratoire Luc Olivier, store residual serum from samples collected for diagnostic purposes (no extra sample is collected). For samples from patients under 20 years of age, we collaborate with partner laboratories (Bruyland, Anacura, Somedi, LBS, Baudouin, Synlab).
Sample collection has started as soon as approval of the ethics committee was obtained (March 30, 2020). Every 3 weeks samples are set aside for 1 week, at least until the end of June 2020. The total number of samples per collection period is, for all labs together, 4000 for the first collection and decreases over the weeks. Collection is stratified: 400 samples per age group of 10 years; the same number of men and women per age group; per age group half of Flanders and the other half of Wallonia and Brussels.
Exclusion of hospital samples and exclusion of duplicates takes place in the laboratories, which also store the samples temporarily until funding and a test are available to determine antibodies against SARS-CoV-2, on behalf of the CEV, UAntwerpen.
Samples are made available unlinked and anonymously, with the only data being: Unique code; Age in years; Collection date; Sex; Postal Code.
Analysis of antibodies against SARS-CoV-2: proportions of seropostive, seronegative or equivalent according to the test used (IgG ELISA, Eurimmune)
Determining overall susceptibility and differences per age group, region and gender by means of suitable methods such as univariate analysis and logistic regression; comparison of these proportions between the different collection periods.
- First sample collection: Result based on 3910 samples: 3% of the samples were sero-positive. Seroprevalence was similar across age groups.
- Second sample collection: Results based on 3397 samples: 6,2% of the samples were sero-positive. The increase was more pronounced in the 20-29 and 80+ age group.
- Third sample collection: Results based on 3253 samples: 6,9% of the samples were sero-positive. This increase is minimal and can be attributed to coincidence variation.