Seroprevalence of anti-SARS-CoV-2 antibodies in women attending antenatal care in eastern Ethiopia

Information on the cumulative incidence of SARS-CoV-2 in East Africa is scarce. We conducted serosurveillance of anti-SARS-CoV-2 antibodies among pregnant women attending their first antenatal care visit in three health facilities in eastern Ethiopia. We collected questionnaire data and a blood sample from 3,312 pregnant women between April 1, 2020 and March 31, 2021 at health facilities in Haramaya, Awoday and Harar. We selected 1,447 blood samples at random and assayed these for anti-SARS-CoV-2 antibodies at Hararghe Health Research laboratory using WANTAI SARS-CoV-2 Rapid Test for total immunoglobulin. Temporal trends in seroprevalence were analysed with a X2 test for trend and multivariable binomial regression. Among 1,447 sera tested, 83 were positive for anti-SARS-CoV-2 antibodies giving a crude seroprevalence of 5.7% (95% CI 4.6%, 7.0%). Of 160 samples tested in April-May, 2020, none was seropositive; the first seropositive sample was identified in June and seroprevalence rose steadily thereafter (X2 test for trend;p=0.003) reaching a peak of 11.8% in February, 2021. In the multivariable model, seroprevalence was approximately 3% higher in first-trimester mothers compared to later presentations, and rose by 0.75% (95% CI 0.31%, 1.20%) per month of calendar time. This clinical convenience sample illustrates the dynamic of the SARS-CoV-2 epidemic in young adults in eastern Ethiopia; infection was rare before June 2020 but it spread in a linear fashion thereafter, rather than following intermittent waves, and reached 10% by the beginning of 2021. After one year of surveillance, most pregnant mothers remained susceptible.


Introduction
In Ethiopia, the first case of COVID-19 was reported on 13 March 2020. By the end of March 56 2021 there were 206,589 reports of COVID-19 infection and 2,865 coronavirus-related deaths. In 57 a country with an estimated population, in 2019, of 112 million this represents a cumulative 58 incidence of SARS-CoV-2 infection of only 0.2% after a full year of transmission. Many cases 59 of COVID-19 present with mild symptoms and, in Ethiopia, three quarters of PCR-positive cases 60 have no symptoms [1,2]. Access to PCR testing in Ethiopia is also sparse. Monitoring the 61 epidemic by detecting symptomatic cases is, therefore, highly insensitive. In these 62 circumstances, seroprevalence of anti-SARS-CoV-2 antibodies can provide a more accurate 63 estimator of cumulative incidence. Undertaking community serosampling during the pandemic is 64 difficult when travel and household access are constrained by control measures. Expectant 65 mothers, however, are likely to continue to seek health services throughout the pandemic and 66 they can be used as a continuously-available proxy population to estimate the cumulative 67 incidence among young adults [3][4][5]. In addition, serological surveillance is simple to implement 68 at ANC clinic visits because anti-SARS-CoV-2 antibodies can be assayed in the residual blood 69 volumes of routine samples collected for clinical screening for anemia and maternal infectious 70 diseases. 71 Planning and provision of health care during a major epidemic like COVID-19 pose substantial 72 logistical and clinical challenges. Information on the shape of the epidemic curve is critical to 73 inform public health responses. The dynamics of seroprevalence reflect the epidemic curve and 74 can provide an estimate of the effective reproduction number. Seroprevalence also indicates the 75 likelihood of approaching transmission control through population immunity. This study aimed 76 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Among the population sample tested, the mean (SD) age was 23.9 (4.7) years and ages ranged 94 from 15 to 45 years. The mean (SD) number of children per mother was 1.5 (1.8). The median 95 (IQR) gestational age at the first antenatal visit was 20 (13-28) weeks. Only 51 (3.5%) had 96 COVID-19 symptoms at the time of sampling and 8 (<1%) had a history of comorbidity, given 97 as chronic liver, renal, cardiovascular or 'other' disease. Respiratory diseases, chronic 98 neurological disease, diabetes mellitus, and cancer were not reported by any participant (Table  99 2). 100

Seroprevalence of SARS-CoV-2 antibodies 105
Of 1,447 samples tested, 83 (5.7%, 95% CI 4.6, 7.0%) were positive for anti-SARS-CoV-2 106 antibodies. The first seropositive sample was identified on June 11, 2020, and seroprevalence 107 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Seroprevalence also varied significantly by trimester of pregnancy and co-morbidity but not by 114 clinic, residence or COVID-19 symptoms (Table 2). Given the linear growth in seroprevalence 115 ( Figure 1) and better model fit based on Bayesian information criterion, we modelled prevalence 116 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

122
The study provides a simple description of the dynamic of SARS-CoV-2 epidemic in an area 123 where reliable data are extremely rare. In a population of attendees at ante-natal clinics in three 124 sites in eastern Ethiopia, antibodies against SARS-CoV-2 first appeared in June 2020 and 125 seroprevalence rose steadily month on month reaching approximately 10% at the beginning of 126 2021. Although the point estimate for March 2021 is substantially lower, the data as a whole 127 evince a strong linear trend and this single estimate is most likely to have deviated from the 128 general direction by chance. If these results are reliable, they indicate that the epidemic is 129 progressing here at a considerably lower rate than in other settings in East Africa and that the 130 greater majority of the population remains uninfected, suggesting that the epidemic is still at an 131 early stage. 132 The is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 21, 2021. Kenya, had a seroprevalence of 44% in August 2020; those in two rural hospitals had 182 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 21, 2021. ; https://doi.org/10.1101/2021.06.17.21257323 doi: medRxiv preprint seroprevalence of 12-13% in November 2020 [27]. Finally, in Addis Ababa seroprevalence, 183 estimated in May 2020, was 3.0% [28]. Although all these studies used different laboratory 184 assays and varied statistical adjustments, collectively, they suggest that transmission in eastern 185 Ethiopia began later than in much of the rest of the region, including the state capital, and has 186 progressed more slowly. 187 In summary if seroprevalence is a reliable indicator of cumulative incidence, SARS-CoV-2 188 infection is spreading slowly but steadily in eastern Ethiopia. This contrasts sharply with the 189 recurrent waves of PCR-positive infections apparent in the national surveillance system. One 190 year after the start of the epidemic approximately 10% of women attending ante-natal clinics are 191 seropositive implying that the COVID-19 epidemic is still at an early stage in eastern Ethiopia. 192

Study area and period 194
The surveillance was conducted between April 1, 2020 and Somali region, and Dire Dawa City. It is one of the ten regional centres designated by the 199 Federal Ministry of Health to manage the COVID-19 epidemic. Haramaya Hospital was rapidly 200 designated a COVID-19 treatment facility and women seeking ANC services were therefore 201 referred to Aweday Health Centre after April 16, 2020. 202 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Statistical analysis 235
We used STATA version 16.0 for statistical analysis. We estimated unadjusted seroprevalence of 236 SARS-CoV-2 IgG antibody with a 95% confidence interval (CI). We did not make adjustment 237 for the test performance characteristics because the manufacturer's validation assay, found very 238 high sensitivity and specificity. We examined the univariate association between individual 239 characteristics and seropositivity using χ 2 and multivariable associations using binomial 240 regression. The trend in seropositivity with time was tested with a χ 2 test for trend and in the 241 multivariable model. 242

Ethical consideration 243
The study was confined to residual clinical blood sample testing and anonymized questionnaire 244 data. It was conducted as part of a public health surveillance, with the approval of the director of 245 each of the three health facilities and the data were made available to relevant bodies including 246 the Regional Health Bureau (Harari and Oromia) and the Ethiopian Public Health Institute 247 (EPHI). Ethical clearance was secured from Institutional Health Research Ethical Review 248 Committee of the College of Health and Medical Sciences, Haramaya University, Ethiopia. 249

Data availability 250
Data is available in the following link and can be requested using the form in the link. give critical feedback. All authors approved the submission of the manuscript to the journal. 261 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 21, 2021. ; https://doi.org/10.1101/2021.06.17.21257323 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 21, 2021. ; https://doi.org/10.1101/2021.06.17.21257323 doi: medRxiv preprint