This analysis uses data longitudinally collected during the COVID-19 pandemic among participants of the Genomes for life (GCAT) cohort in Catalonia. The GCAT cohort includes mainly middle-aged participants who are residents in Catalonia, and their recruitment started in 2014 (11). Most participants were enrolled from blood donors invited through the Blood and Tissue Bank and are regularly followed-up. We contacted GCAT eligible participants just after the strict first confinement period in 2020 and almost a year later in 2021 after COVID-19 vaccine administration began in Spain. Participants were contacted via email or telephone and asked to respond to a questionnaire (online or via telephone) and provide a blood sample both in 2020 and 2021 follow-up. Blood sampling in 2021 was offered to all 2020 participants with a seropositive or an undetermined serostatus (response rate, 47.15%); and to a random sample of 2020 seronegative participants (response rate, 44.21%). People were aware of their 2020 serology results. Additionally to the serology and questionnaires data, results on viral detection tests (SARS-CoV-2 polymerase chain reaction (PCR) and rapid antigen-testing) and immunization data were collected from Electronic Health Records of the Epidemiological Surveillance Emergency Service of Catalonia of the Department of Health. The timeline of serology assessments in our population along with the dates of vaccination administration and the positive testing are all presented in supplementary figure 1.
Blood samples collected at both time points were processed within 24 h of collection and were analyzed at the ISGlobal Immunology laboratory in Barcelona. The levels [median fluorescence intensity (MFI)] of IgG, IgM and IgA were assessed by high-throughput multiplex quantitative suspension array technology against a panel of 5 SARS-CoV-2 antigens: the spike full length protein (S) and the receptor binding domain (RBD) produced at IDIBAPs (both fused with C-terminal 6xHis and StrepTag purification sequences and purified from supernatant of lentiviral-transduced CHO-S cells cultured under a fed-batch system) (12), the sub-region S2 (purchased from SinoBiological), the nucleocapsid (N) full length (FL) and the specific C-terminal (Ct) region produced at ISGlobal (both expressed in E. coli and His tag-purified (13). In addition, 4 RBDs from different VoCs, produced at CRG (expressed in Expi294 and His tag purified) were used (with RBD amino acid changes from Wuhan-Hu-1 in parentheses): Alpha (N501Y) (UK, September 2020), Beta (K417N, E484K, N501Y) (South Africa, September 2020), Gamma (K417T, E484K, N501Y) (Brazil, December 2020), Delta (L452R, T478K) (India, December 2020). Αlpha was the major variant circulating in Europe until early Summer 2021 and was then quickly displaced by the Delta variant, whereas Beta and Gamma accounted only for some cases among those samples sequenced by authorities across Europe at that time (14).
Assay performance was previously established as 100% specificity and 95.78% sensitivity for seropositivity 14 days after symptoms onset (15). Antigen-coupled microspheres were added to a 384-well μClear® flat bottom plate (Greiner Bio-One, Frickenhausen, Germany) in multiplex (2000 microspheres per analyte per well) in a volume of 90 μL of Luminex Buffer (1% BSA, 0.05% Tween 20, 0.05% sodium azide in PBS) using 384 channels Integra Viaflo semi-automatic device (96/384, 384 channel pipette). Hyperimmune pools were used as positive controls prepared at two-fold, 8 serial dilutions from 1:12.5. Pre-pandemic samples were used as negative controls to estimate the cut-off of seropositivity. Ten microliters of each dilution of the positive control, negative controls and test samples (prediluted 1:50 in 96 round-bottom well plates), were added to a 384-well plate using Assist Plus Integra device with 12 channels Voyager pipette (final test sample dilution of 1:500 for all isotypes, and a second dilution at 1:5000 for IgG to assess response to S proteins in vaccinated subjects avoiding signal saturation). To quantify IgM and IgA, test samples and controls were pre-treated with anti-Human IgG (Gullsorb) at 1:10 dilution, to avoid IgG interferences. Technical blanks consisting of Luminex Buffer and microspheres without samples were added in 4 wells to control for non-specific signals. Plates were incubated for 1 h at room temperature in agitation (Titramax 1000) at 900 rpm and protected from light. Then, the plates were washed three times with 200 μL/well of PBS-T (0.05% Tween 20 in PBS), using BioTek 405 TS (384-well format). Twenty-five microliter of goat anti-human IgG-phycoerythrin (PE) (GTIG-001, Moss Bio) diluted 1:400, goat anti-human IgA-PE (GTIA- 001, Moss Bio) 1:200, or goat anti-human IgM-PE (GTIM-001, Moss Bio) 1:200 in Luminex buffer were added to each well and incubated for 30 min as before. Plates were washed and microspheres resuspended with 80 μL of Luminex Buffer, covered with an adhesive film and sonicated 20 s on sonicator bath platform, before acquisition on the Flexmap 3D reader. At least 50 microspheres per analyte and per well were acquired, and MFIs were reported for each isotype-antigen combination. Assay positivity cut-offs specific for each isotype antigen combination were calculated as 10 to the mean plus 3 standard deviations of log10-transformed MFI of 128 pre-pandemic controls. Results were defined as undetermined when the MFI levels for a given isotype-antigen combination were between the positivity threshold and an upper limit at 10 to the mean plus 4.5 standard deviations of the log10-transformed MFIs of pre-pandemic samples, and no other isotype-antigen combination was above the positivity cut-off. We defined overall serostatus, by isotype and by antigen. Spearman correlation between RBD log10 transformed responses of different VoC was generally high (Supplementary table 1). Among individuals who tested negative/undetermined for antibodies to RBD Wuhan, five were RBD positive for at least another VoC (Supplementary table 2). Among RBD Wuhan seropositive, there were 215 people with no detectable antibodies against RBD to at least one other VoC and 34 people with no detectable antibodies to the RBD of all other VoC. Thus, we based the overall serostatus only on RBD Wuhan responses.
During the course of this study, the vaccines approved for use in Spain were the Comirnaty (BNT162b2, mRNA, BioNTech-Pfizer, Mainz, Germany/New York City, United States (US); the Spikevax (mRNA-1273, Moderna, Cambridge, US); the Vaxzevria (ChAdOx1 nCoV-19, Oxford–AstraZeneca) and the Janssen COVID-19 vaccine (Ad26.COV2.S, Johnson & Johnson–Janssen). The first doses (January 2021) were distributed among some of the most vulnerable groups, including residents and personnel working in retirement homes, front-line healthcare workers, highly dependent people, and seniors aged 80 and older. Subsequently, essential workers and then adults in descending age order were vaccinated. The government restricted vaccination with Vaxzevria to specific age groups (first upper threshold of 55, then 65, and finally was restricted for people 60-69 years old) and offered the Comirnaty or the Spikevax vaccine as 2nd dose for those <60 years old. The vaccines were offered on a voluntary basis to all individuals.
In this study, we used electronic registries to identify the number of doses, date of administration and trade names of vaccines for each study participant. We defined participants as non-vaccinated, partially vaccinated (when they had received one dose for those vaccines that have a two-dose regimen and had not acquired infection previously) and fully vaccinated (when they had received two doses for those vaccines that have a two-dose regimen or one dose in those previously infected with SARS-CoV-2 [recommended after six months following infection] or one dose for vaccines that have a one-dose regimen).
Performance and results on viral detection tests (PCR or antigen test) were either self-reported or identified by the use of the official repository of tests for SARS-CoV-2 of the Epidemiological Surveillance Emergency Service of Catalonia of the Department of Health (16). We combined two sources of information because each had its own limitations. Our strategy for detection of participants with evidence of prior infection up to the 2021 serology included i) previous positive viral detection test, ii) seropositivity in 2020 sample (pre-vaccination), iii) seropositivity in 2021 sample among non-vaccinated, iv) seropositivity to N-antigen in 2021 sample of those vaccinated, since the available vaccines do not contain or produce N-antigen.
Descriptive analyses of the study population characteristics and comparisons between groups were conducted. In all analyses we used antibody levels log10 transformed, due to their skewed distribution. Differences in antibody levels by vaccine type or by vaccination and/or infection were examined using one-way ANOVA and pairwise comparisons were performed using Tukey post hoc-test. Previously infected non-vaccinated individuals were classified for each isotype-antigen combination, as sustainers when the ratio of antibody levels between the two visits ≥1 or decayers when the ratio of antibody levels between the two visits <1 following previous methodology (10,21). In increasers and decayers, linear mixed-effects models with random intercept and random slope were used to evaluate the trend in antibody levels between the two serology time points, we estimated fold-change one year after infection separately in sustainers and decayers. Generalized additive models were used to explore the shape of the relationship between days since vaccination and antibody levels among vaccinated people stratified by infection, models were done separate for participants with 1 dose or 2 doses at the time of serology. Among vaccinated people, univariable and multivariable stepwise linear regression models were fit to determine the effect of several variables on the antibody levels. We excluded only people vaccinated with Janssen COVID-19 vaccine because they were few and eligible only for one dose at that time. We evaluated all the determinants in univariable models and also built a final model for each isotype-antigen combination using forward stepwise regression models including all the described determinants considering 0.1 the significance level for addition to the model and 0.2 for removal from the model. We report associations as a fold-change with 95% confidence intervals (FC 95% CI) obtained by 10^beta. We considered the following variables as determinants: age (<60 or ≥60 years old), sex (male, female), highest attained educational level (primary or less, secondary, university), area-based socioeconomic status according to area of residency (in quantiles), current smoking status (smoker, non-smoker), daily alcohol consumption (yes, no), BMI status (obese, overweight, versus normal/underweight), SARS-CoV-2 infection, vaccine type and self-reported information on chronic diseases (a disease in the last 6 months that required visit to the doctor or medical treatment) including cardiovascular diseases (hypertension, heart attack, angina pectoris), mental health diseases/addictions (yes, no) and immune related diseases (rheumatoid arthritis, other autoimmune diseases, HIV or other immunodeficiency, asthma). We adjusted all models a priori for time since the last vaccine dose (<1 month, 1-2 months, 2-3 months, 3-4 months, >4 months) and number of doses (one or two). Analysis was repeated by vaccine type. Participants with missing covariates were excluded from the final analysis models. We performed all statistical analyses using Stata/SE (version 16; StataCorp LLC.).