Evolution of seroprevalence to SARS‐CoV‐2 in blood donors in Sarajevo Canton, Federation of Bosnia and Herzegovina: Cross‐sectional and longitudinal studies

Abstract Background Sarajevo Canton in the Federation of Bosnia and Herzegovina has recorded several waves of high SARS‐CoV‐2 transmission and has struggled to reach adequate vaccination coverage. We describe the evolution of infection‐ and vaccine‐induced SARS‐CoV‐2 antibody response and persistence. Methods We conducted repeated cross‐sectional analyses of blood donors aged 18–65 years in Sarajevo Canton in November–December 2020 and 2021. We analyzed serum samples for anti‐nucleocapsid (anti‐N) and anti‐spike (anti‐S) antibodies. To assess immune durability, we conducted longitudinal analyses of seropositive participants at 6 and 12 months. Results One thousand fifteen participants were included in Phase 1 (November–December 2020) and 1152 in Phase 2 (November–December 2021). Seroprevalence increased significantly from 19.2% (95% CI: 17.2%–21.4%) in Phase 1 to 91.6% (95% CI: 89.8%–93.1%) in Phase 2. Anti‐S IgG titers were significantly higher among vaccinated (58.5%) than unvaccinated infected participants across vaccine products (p < 0.001), though highest among those who received an mRNA vaccine. At 6 months, 78/82 (95.1%) participants maintained anti‐spike seropositivity; at 12 months, 58/58 (100.0%) participants were seropositive, and 33 (56.9%) had completed the primary vaccine series within 6 months. Among 11 unvaccinated participants who were not re‐infected at 12 months, anti‐S IgG declined from median 770.1 (IQR 615.0–1321.7) to 290.8 (IQR 175.7–400.3). Anti‐N IgG antibodies waned earlier, from 35.4% seropositive at 6 months to 24.1% at 12 months. Conclusions SARS‐CoV‐2 seroprevalence increased significantly over 12 months from end of 2020 to end of 2021. Although individuals with previous infection may have residual protection, COVID‐19 vaccination is vital to strengthening population immunity.

Serosurveys provide critical estimates of population immunity, particularly in areas with suboptimal disease surveillance and vaccine coverage. 7 Asymptomatic or mild disease is common, which leads to a high prevalence of under-detection. 8,9 Although knowledge of the virological, epidemiological, and clinical characteristics of SARS-CoV-2 has progressed considerably since the beginning of the pandemic, few studies have been conducted in the Western Balkan region.
Using an adapted WHO UNITY study seroepidemiological investigation protocol, 10 we aimed to estimate the seroprevalence of anti-SARS-CoV-2 antibodies in Sarajevo Canton by conducting repeated cross-sectional analyses among blood donors (aged 18-65 years) at 12-month intervals from end of 2020 to end of 2021; to assess immune durability at 6-and 12-month intervals among those who were seropositive at end of 2020; and to describe the antibody response among blood donors with infectionand/or vaccine-induced immunity, including across vaccine products.
This will support improved understanding of population immunity and provide evidence to support informed public health decision making.

| Study design and participants
We conducted repeated cross-sectional analyses among blood donors aged 18 to 65 years who attended the Institute of Transfusion Medicine of the FBiH at end of 2020 (Phase 1) and end of 2021 (Phase 2) to estimate the seroprevalence of anti-SARS-CoV-2 antibodies in Sarajevo Canton. We aligned methods with WHO standardized protocol under the Unity Studies initiative. 10 In the cross-sectional analyses, blood donors were invited to participate in the study, irrespective of prior SARS-CoV-2 infection (Figure 1). Nonresidents of Sarajevo Canton and those experiencing COVID-19 symptoms, or who reported contact with a confirmed COVID-19 case within 14 days prior, were excluded.
In the longitudinal analysis, seropositive participants in Phase 1 were contacted via telephone and invited to participate in follow-up assessments at approximately 6-and 12 months after baseline to assess immune durability. Participants were excluded if they refused to participate or were no longer residents of Sarajevo Canton.
Partially vaccinated participants, and those who had completed the primary series more than 6 months prior or had received a third booster dose were included in overall seroprevalence estimates but excluded from other analyses that aimed to estimate the effect of vaccination on population immunity.

| History of COVID-19 and vaccination
After providing written consent, all eligible participants completed a questionnaire regarding their history of suspected or confirmed COVID-19 infection within the last 9 months (Phase 1) or 3 months (Phase 2), including whether they had experienced COVID-19 symptoms, had received a diagnosis of COVID-19, or had been hospitalized due to COVID-19. Participants also provided their vaccination status (Phase 2) via verbal report or vaccination card. We determined individuals to be fully vaccinated against COVID-19 if they had completed the primary vaccine series at least 14 days but no more than 180 days prior to serological testing. Vaccines were stratified by platform:

| Serological testing
Trained laboratory staff at the Institute for Transfusion Medicine of the FBiH collected venous blood samples from participants and conducted laboratory testing. Plasma samples (500 μL aliquot) of seropositive participants were stored at À30 C and archived for possible additional testing for up to 2 years, as described on the informed consent forms signed by participants. Anti-SARS-CoV-2 antibodies were detected by chemiluminescent microparticle immunoassays (CMIA) using the Abbot Laboratories SARS-CoV-2 IgG I and II kits and ARCHITECT i2000sr system.
In Phase 1, serum samples were assessed for anti-nucleocapsid (anti-N); 1.4 AU/mL or greater was considered seropositive for anti-N.
In Phase 2, serum samples were assessed for the receptor binding domain (RBD) of the S1 subunit of the spike protein of SARS-CoV-2 in serum and plasma (anti-S); 50.0 AU/mL or greater was considered seropositive for anti-S. F I G U R E 1 Participant flow diagram. Participants were excluded due to refusal to participate or residence outside Sarajevo Canton.
In the longitudinal analysis, anti-N seropositive participants at baseline were contacted via telephone and invited to participate in follow-up assessment at 6-and 12 months for anti-N and anti-S seropositivity. Using stored aliquots, we retroactively assessed anti-S antibodies among participants who were seropositive for anti-N at baseline and included in the 6-month follow-up.

| Statistical analysis
Participants were stratified by age (18-29, 30-39, and 40-65 years) and sex. The crude seroprevalence estimates and 95% confidence intervals were adjusted using direct standardization for the age and to 1500 blood donations per month. Assuming a seroprevalence of 5%-50% during the study period and a 95% confidence interval, inclusion of 300-500 blood donors per age stratum would allow a 2%-6% margin of error for seroprevalence estimates. All statistical analyses were conducted in R 4.2.0.   The majority of seropositive participants in Phase 2 were vaccinated, with vaccination increasing significantly by age (p = 0.001) ( Figure S2). Among 364 unvaccinated participants, 283 (77.7%) were seropositive in Phase 2. There were no statistically significant differences in vaccination between men and women ( p = 0.347). Fully vaccinated participants recorded significantly higher anti-S titer values than unvaccinated infected participants, regardless of the vaccine product ( Figure 2). However, no statistically significant differences were detected between vaccine products that shared the same platform. Based on pairwise comparison of each product, anti-S IgG titers were highest among participants who had received mRNA vaccines, followed by vector-based vaccines, and inactivated vaccines (p < 0.05).  had completed the primary series more than 6 months prior (Table 3 and Figure 3). Vaccination status was reported verbally.   Figure S1).

| Longitudinal analysis
The estimated seroprevalence in Phase 1 (17.7%) was significantly higher than reported cumulative incidence based on testing (5.7% by December 31, 2020), 6 which was not unexpected and has been observed elsewhere. For example, a 2021 systematic review estimated the seroprevalence to be 1.5 to 717 times higher than the cumulative reported case incidence. 11 Disparities might be explained by testing constraints or the prevalence of asymptomatic, atypical, or pauci-symptomatic cases. 12 In Phase 1 of our cross-sectional analysis, half of seropositive participants reported no history of symptoms associated with COVID-19, and majority were younger than 40 years. This is consistent with previous studies, which found younger age groups to be more likely to be asymptomatic carriers of SARS-CoV-2, less likely to comply with PHSM, and less likely to seek medical attention if clinical symptoms arise. 13,14 Despite the fact that both phases were conducted following the high SARS-CoV-2 transmission wave, participants in Phase 1 reported more COVID-19 infection history than participants in Phase 2, perhaps as a result of the longer timeframe for reporting history of symptoms.
Vaccination against SARS-CoV-2 significantly contributed to high seroprevalence during Phase 2 of our study. Global studies have demonstrated the critical role of vaccination in achieving population immunity to SARS-CoV-2, primarily to protect against severe disease. 12,15 A study conducted in the United States among blood donors estimated that seroprevalence increased from 3.5% to 11.5% between July and December 2020 and then to 83.3% by May 2021 as a result of infection-and vaccine-induced antibody response, with F I G U R E 3 Anti-N and anti-S IgG antibody response at baseline, 6 months, and 12 months among primarily infected, re-infected, and vaccinated blood donors, Sarajevo Canton, BiH, Nov/Dec 2020-Nov 2021/Jan 2022. Primarily infected: unvaccinated participants whose anti-S and anti-N IgG titer values decreased over the follow-up period; re-infected: unvaccinated participants whose anti-S and/or anti-N IgG titer values increased over the follow-up period; vaccinated: participants who had completed the primary vaccine series. At 6 months, analysis included 64 primarily infected and 16 re-infected participants; two partially vaccinated participants were excluded. At 12 months, analysis included 11 primarily infected, 12 re-infected since baseline, and 30 vaccinated participants ( the majority of the seropositivity attributable to vaccination. 16 Similarly, a study in Geneva, Switzerland, conducted in July 2021, found a population seroprevalence of 66.1%, where 36.2% had developed antibodies due to vaccination. 17 We found that fully vaccinated participants (including those who might have previously been infected) recorded significantly higher anti-S titer values than unvaccinated, infected participants, regardless of the vaccine product, though titers varied across vaccine platforms.
We also observed a higher antibody response among vaccinated per- however, the overall risk of re-infection was lowest among those who were vaccinated following infection (hybrid immunity) during periods of Delta predominance. 19,20 Anti-S IgG titers among seropositive participants in our study remained stable over 12 months. This finding is consistent with previous research among confirmed COVID-19 cases from 6 to 12 months following infection, though antibody response following infection was highly heterogeneous between individuals. [21][22][23][24][25][26][27] Previous studies also found anti-N antibody response wanes earlier, becoming undetectable in most cases by 5 to 7 months. 28,29 Although the duration of immunity may depend on the type of vaccine, anti-S antibodies remain detectable at least 6 to 8 months following vaccination. 30 32 Few participants in our study reported severe disease, with hospitalization reported by just three participants in Phase 1 and one participant in Phase 2, so it was not possible to assess an association with antibody response. Finally, the manufacturer's cut-off threshold of the employed test may lead to the underestimating of seroprevalence, as suggested in certain peer-reviewed papers. 33,34 This study has many strengths. Our study addresses an important gap in COVID-19 surveillance in Bosnia and Herzegovina and contributes evidence from the Western Balkans, a region that has reported among the highest mortality associated with COVID-19. 35 We used validated tests and adjusted for the age and sex distribution in the population of Sarajevo Canton. We also estimated durability of two types of antibodies (anti-N and anti-S), which are known to have different periods of detectability. [36][37][38][39] We also monitored the evolution of seroprevalence to SARS-CoV-2 following its introduction to a welldefined geographic area and present vaccination data across three COVID-19 vaccine platforms. Finally, use of the standardized WHO UNITY protocol allows for comparison and synthesis of our study results with other studies globally.
This study was conducted prior to the widespread circulation of the SARS-CoV-2 variant of concern, Omicron ( Figure

ACKNOWLEDGMENTS
We would like to thank laboratory technicians who took part in the study and all participants who voluntary agreed to participate in the study. We would also like to thank Alina Guseinova, Gudrun Freidl, Yuster Ronoh and other colleagues at the World Health Organization who supported this study.

CONFLICT OF INTEREST STATEMENT
VIS has served as a consultant for WHO Country Office in Bosnia and Herzegovina. Co-authors report no conflicts of interest.

PEER REVIEW
The peer review history for this article is available at https://www. webofscience.com/api/gateway/wos/peer-review/10.1111/irv.

DATA AVAILABILITY STATEMENT
De-identified data and R script will be made available upon request.

ETHICS STATEMENT
The protocols for each phase of this study were reviewed and