key: cord-0721231-ric9hh2k
authors: Nkuba, Antoine N; Makiala, Sheila M; Guichet, Emilande; Tshiminyi, Paul M; Bazitama, Yannick M; Yambayamba, Marc K; Kazenza, Benito M; Kabeya, Trésor M; Matungulu, Elysee B; Baketana, Lionel K; Mitongo, Naomi M; Thaurignac, Guillaume; Leendertz, Fabian H; Vanlerberghe, Veerle; Pelloquin, Raphaël; Etard, Jean-François; Maman, David; Mbala, Placide K; Ayouba, Ahidjo; Peeters, Martine; Muyembe, Jean-Jacques T; Delaporte, Eric; Ahuka, Steve M
title: High prevalence of anti-SARS-CoV-2 antibodies after the first wave of COVID-19 in Kinshasa, Democratic Republic of the Congo: results of a cross-sectional household-based survey
date: 2021-06-05
journal: Clin Infect Dis
DOI: 10.1093/cid/ciab515
sha: ade411bf230f41b794154458b72c7e9f098d546c
doc_id: 721231
cord_uid: ric9hh2k

BACKGROUND: On October, 2020, after the first wave of COVID-19, only 8290 confirmed cases were reported in Kinshasa, Democratic Republic of the Congo, but the real prevalence remains unknown. To guide public health policies, we aimed to describe the prevalence of SARS-CoV-2 IgG antibodies in the general population in Kinshasa. METHODS: We conducted a cross-sectional, household-based serosurvey between October 22, 2020, and November 8, 2020. Participants were interviewed at home and tested for antibodies against SARS-CoV-2 spike and nucleocapsid proteins in a Luminex based assay. A positive serology was defined as a sample that reacted with both SARS-CoV-2 proteins (100% sensitivity, 99.7% specificity). The overall weighted, age-standardized prevalence was estimated and the infection-to-case ratio was calculated to determine the proportion of undiagnosed SARS-CoV-2 infections. RESULTS: A total of 1233 participants from 292 households were included (mean age, 32.4 years; 764 [61.2%] were women). The overall weighted, age-standardized SARS-CoV-2 seroprevalence was 16.6% (95% CI 14.0-19.5). The estimated infection-to-case ratio was 292:1. Prevalence was higher among participants ≥ 40 years than among those ˂18 years (21.2% vs 14.9%, respectively; p˂0.05). It was also higher in participants who reported hospitalization than among those who did not (29.8% vs 16.0%, respectively; p˂0.05). However, differences were not significant in the multivariate model (p=0.1). CONCLUSION: The prevalence of SARS-CoV-2 is much higher than the number of COVID-19 cases reported. These results justify the organization of a sequential series of serosurveys by public health authorities to adapt response measures to the dynamics of the pandemic.

identified in December 2019 in China, but has spread extremely fast worldwide, including in Africa [1] . In the Democratic Republic of Congo (DRC), the first confirmed COVID-19 case was reported on March 10, 2020, in the capital city of Kinshasa, in a Congolese traveller who lived in Europe and had returned to the DRC. After the first cases, the Congolese government rapidly declared a state of emergency and set up a national multi-sectoral national committee to design strategies to address the pandemic [1] . Among the public health measures taken to control the spread of the virus, the national lockdown was first imposed in Kinshasa and then across the entire country, flights from COVID-19-infected countries were suspended, schools and universities were closed, gatherings of more than 20 people were banned, and there was an obligation to wear masks in public areas [2] .

On October 19, 2020 , after the first epidemic wave (March to July), the DRC reported 11 078 confirmed COVID-19 cases with 303 deaths. The capital city of Kinshasa represented 74% (8290) of all notified cases across the country [3] . COVID-19 reported cases in the DRC were much lower than predicted by many researchers, For example, Walker et al suggested that 70 million Africans could be infected by SARS-CoV-2, with more than 3 million deaths [4] . Furthermore, Wells et al estimated that there would be 76 213 155 infections and 319 441 deaths in the absence of physical distancing and any public health measures in the DRC [5] .

Several factors may explain this glaring difference between the prediction and the number of COVID19 reported cases. First, the limited capacity to test for SARS-CoV-2, especially in the early stage of the epidemic, but gradually the daily capacity of SARS-CoV-2 polymerase chain reaction (PCR) testing in Kinshasa increased and reached 1000, which is still insufficient. However, there was no saturation of hospital reception capacities or unexplained high mortality during the same period, even if mortality may have been underestimated. Some of the potential factors that could explain this difference are the age pyramid with a younger population potentially resulting in a greater A c c e p t e d M a n u s c r i p t 5 number of asymptomatic cases, pre-existing immunity due to possible cross-reaction with other tropical infectious diseases or other coronaviruses, environmental factors, and early implementation of measures to control the disease [6, 7] .

In the African context, and following the recommendations of the WHO, [8] population-based serosurveillance is important to complete data on the reported cases of SARS-CoV-2 infection in order to assess the real extent of the epidemic and to enable decision-makers to adjust public health response measures. Several serological surveys have been carried out in Africa and have shown a high variability of seroprevalence to SARS-CoV-2 across countries, but most were performed in specific population groups, such as blood donors, health care workers or other high-risk populations [9] [10] [11] . Apart from epidemiological factors, the difference in reported seroprevalence might be explained by the different types of serological assays used, especially those that are designed to detect a single IgG antibody, and to their target populations. We recently reported on the challenges of SARS-CoV-2 seroprevalence studies done in African countries with commercial tests validated in Europe, the USA, or Asia [12] .

We aimed to describe the prevalence of IgG antibodies to SARS-CoV-2 in the general population of Kinshasa in order to understand to what extent the virus has spread after the first epidemic wave.

The 2020 Appui à la Riposte Africaine à l'épidémie COVID19 [Support for the African Response to COVID-19] (ARIACOV) survey was a household-based seroprevalence survey conducted between October 22, 2020 and November 8, 2020, in Kinshasa. The sampling frame used the health divisions of the city. Kinshasa is divided into 35 health zones, which are divided in turn into 380 health areas, with an estimated total population of 12 117 417 inhabitants (Système National d'Information Sanitaire [National Health Information System]). A health zone is defined as an operational unit, A c c e p t e d M a n u s c r i p t 6 which supports 100,000 to 150,000 inhabitants and is delimited taking into account the geographical, cultural and economic accessibility of the population. It could be a geographical space contained within the limits of a territory or an administrative commune comprising a population of approximately 100,000 to 250,000 people in urban areas. Kinshasa is divided in 35 health zones for 26 administrative communes based on the number of population within each administrative commune (unit). A three-stage design was used to randomly select 292 households (figure S1, appendix p.ii). First, 14 of 35 health zones (divided into two strata corresponding to the eastern and western regions of the city) were selected with a probability to be selected proportional to the number of households in each zone. Then, within each zone, three health areas were randomly selected and finally eight households were selected within each health zone. To balance the groups, all the residents were invited to participate in the study in 50% of households and among the remaining 50%, only the residents aged 18 years and older were invited to participate.

During the study, all individuals with a suspicion of COVID-19 infection were referred for PCR testing and patient care to the COVID-19 reference centre. All staff involved in the study were tested by PCR prior to the survey and followed infection, prevention and controls recommendations. Communitybased mobilization for the survey was performed in a two-step process. The study team met with local leaders and key stakeholders a few weeks prior to the start of the survey and also visited each selected cluster to directly mobilize the community about the survey. Participants were told that the survey was about COVID-19/SARS-CoV-2 and that they would be tested for antibodies if they agreed to participate. Ethics approval was obtained from the Comité d'Ethique de l'Ecole de Santé Publique de Kinshasa (protocol no. ESP/CE/156/2020). All adults and children (≥10 years) were informed about the study objectives and procedures. Adults provided written consent to participate in the study and to be tested for SARS-CoV-2 serology prior to starting the interview. Written parental consent and children ascent when ≥10 years were obtained prior to enrollment of participants ˂18 years.

Venous blood samples (3-5 ml) were collected from eligible participants in a "red-top tube", which did not contain any additives and transported to the National Institute of Biomedical Research. After centrifugation, serum samples were aliquoted and stored at -20°C until laboratory analysis. Presence for antibodies to SARS-CoV-2 was done with a previously developed, highly sensitive and specific Luminex-based assay to simultaneously detect IgG antibodies to two viral antigens, i.e recombinant nucleocapsid (NC) and spike (SP) proteins derived from SARS-CoV-2, as previously described [13] .

Results were expressed as median fluorescence intensity for 100 beads. Cutoff values were determined with receiver operating characteristic curve analysis from a panel of SARS-CoV-2 negative and positive plasma samples consisting of European donors before the COVID-19 pandemic and hospitalized PCR-confirmed patients, respectively [13] . Specificity was validated on a panel of 1197 samples from Africa before COVID-19 (99.7% specificity, Supplementary Table S1). A sample was considered positive for immunoglobulin G (IgG) against SARS-CoV-2 if it reacted simultaneously with NC and SP proteins. As several studies have reported a decrease in antibody levels over time, we considered samples with only one of two antigens above the threshold as "indeterminate" due to the difficulty to discriminate between antibody decline or the lower specificity of single antigen reaction, as often reported in samples from Africa [12, 14] . Samples with median immunofluorescence intensity (MFI) below the cutoff for both antigens were considered as negative.

A c c e p t e d M a n u s c r i p t 8

Statistical analysis was done using Stata 16 (StataCorp, College Station, TX, USA). Data were checked and analyzed using the svyset commands to take into account the survey design. Descriptive statistics were weighted to take into account the selection probability of the cluster sampling procedure and are presented as proportions with their 95% confidence intervals or means with standard deviation. The Pearson's chi-squared test was used to compare categorical descriptive outcomes. The overall prevalence estimate was weighted and age-standardized based on available demographic data [15] . Multivariate logistic models were used to assess the association between positive serology and key risk factors. Likelihood ratio tests were performed to determine the significance of each factor in the model. To estimate the total number of SARS-CoV-2 infections in the population, we multiplied the weighted, age-standardized seroprevalence by the population of Kinshasa at the time of the survey and divided this number by the number of reported COVID-19 cases detected by reverse transcription-PCR on October 19, 2020, to estimate the infection-to-case ratio.

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Among 292 randomly-selected households from 42 clusters, 2400 individuals were eligible; 1607 were present at the time of the survey and 1233 (76.7%) were included in the final analysis ( Figure   1 ; Table 1 ). Of these, 1080 provided sufficient and compliant samples that were tested for anti-SARS-CoV-2 IgG antibodies against spike and nucleocapsid proteins. Mean age participants was 32.4 ± 19.5 years: 461(37.4%) were in the 18-39 age group, 420 (34.1%) in the ≥ 40 age group, and 352 A c c e p t e d M a n u s c r i p t 9 (28.6%) in the 0-17 years age group; 764 (61.2%) were women. Most participants (72.2%) resided in a common yard and 668 (54.2%) did not have access to hand washing devices at home. In total, 750 (60.8%) declared having completed secondary studies and 349 (28.3%) were pupils or students.

Overall, 659 (53.5%) participants were from the eastern part of the city of Kinshasa.

The overall weighted, age-standardized SARS-CoV-2 seroprevalence was 16.6% (95% CI 14.0-19.5) with both anti-IgG against SP and NC proteins. In addition, 17.1% ( A c c e p t e d M a n u s c r i p t 10 and nucleocapsid proteins but no significant differences were observed among those who reported symptoms. We did not observe associations with a single or a combination of clinical symptoms and seropositivity. We also assessed whether hospitalization could be associated with seroprevalence and observed that the proportion of people who admitted to have been hospitalized before the survey and positive for anti-SARS-CoV-2 antibodies (29.8% [95% 17.0-46.8]) was significantly higher compared to those who were not hospitalized (16.0% [95% CI 13.4-19.0]) ( Table 3) . None of the households reported deaths with symptoms related to COVID-19.

We then performed multivariate analysis for all parameters with significant difference of seroprevalence between groups, but no significant association between age, gender and hospitalization with seroprevalence was found ( Table 4 ).

To our knowledge, this is the first serological survey conducted in the general population of Kinshasa after the first wave of the COVID-19 pandemic (March to September, 2020). The overall prevalence of anti-SARS-CoV-2 antibodies was 16.6%. Extrapolation to the entire population showed that around 2.4 million infections occurred between March and October 2020 in contrast to the 8290 PCR confirmed cases reported during the same period. We estimated that most cases went unnoticed, with only one case detected for every 292 infections.

Seroprevalence in the DRC was higher than that reported in India, Brazil, Switzerland and Zambia, [16] [17] [18] [19] but was probably linked to the different age structure or to the early stage of the COVID-19 pandemic, as was the case in Zambia [19] . Several other countries in Africa have reported varying seroprevalences eg, the prevalence was lower in Kenya, Togo and Malawi [9, 10, 20] , but higher in Niger, the Ivory Coast and South Sudan [21] [22] [23] . This variability could be explained first by the fact that most of these studies were carried out in specific populations and second, apart from A c c e p t e d M a n u s c r i p t 11 the Malawi study, all the other studies reported a seroprevalence using positivity against a single antigen (spike or nucleocapsid).

Although we observed a trend among age groups, seroprevalence was not significantly associated with age groups in our survey. Indeed, several studies have reported a different distribution of seroprevalence according to age, whereas others did not observe age-related differences, eg, in India, Brazil and Zambia, seroprevalence was similar between age groups [16, 17, 19] . However, a study conducted in Iran reported that the prevalence of COVID-19/SARS-CoV-2 varied by age group [24] . In Switzerland, seroprevalence was significantly lower among young children (5-9 years) and older people (≥65 years) than for other age groups [18] . Only limited data on the kinetics of antibodies in children are available and most tests were validated on the samples of adults with symptoms. However, antibody titers and kinetics in children exposed to SARS-CoV-2 are most likely similar to those in adults [25] .

On the basis of our pilot study that showed a low agreement between the results of commerciallyavailable antibody detection assays validated in Europe, the USA, or Asia on African samples, we recommend the use of a combination of serological tests, targeting two or more independent antigens in this context [12] . This is even more important with the advent of mass vaccination. Therefore, assuming that seropositivity against spike and nucleocapsid proteins is evidence of true seroconversion, we estimated the seroprevalence of antibodies against SARS-CoV-2 at 16.6%. During the first wave (March to September), the majority of notified PCR confirmed cases came from the western health zone of Kinshasa, but our study showed that seroprevalence was almost similar between the two geographic regions studied, thus suggesting that COVID-19 had spread throughout the entire city despite early government actions. Nevertheless, this high seroprevalence was not accompanied by higher mortality rates or saturation of hospital services. The SARS-CoV-2 virus is known for its ability to be transmitted to all ages, but the risk to develop a severe form increases with age and other risk factors, including obesity among younger individuals [26, 27] . Indeed, the A c c e p t e d M a n u s c r i p t 12 age pyramid in DRC is made up of a large base of young people, with a small top of the elderly, and people aged ≥65 years represent only 3% of the total population [15] .

In many studies, seroprevalence is determined using only one of the two major SARS-CoV-2 antigens [9-11, 16, 28] . This may lead to an overestimation due to cross-reactivity of the spike or nucleocapsid antibodies against other viral antigens [29] or underestimation by antibody waning [14] . Cross-reactivity can occur with common circulating coronaviruses, as well as other viruses such as dengue [29] . In contrast to anti-spike IgG antibodies, which are sustained over time, the half-life of the SARS-CoV-2 anti-nucleocapside IgG response seems to be shorter [14] . Seroprevalence of antibodies against the spike or nucleocapsid only was 17.1% in our study. The overall seroprevalence in Kinshasa combining all positive and indeterminate groups could thus be higher and reach 32.6%. Seroprevalence based on positivity to two different SARS-CoV-2 antigens provides thus most likely minimal estimates, and it is probable that a proportion of the participants with antibodies against a single antigen also represented individuals that had a previous SARS-CoV-2 infection. Nevertheless, most of the population from Kinshasa remains not infected as yet and it is to be hoped that the spread of SARS-CoV-2 can be maintained until the herd immunity threshold is achieved. This estimated threshold is approximately 50% to 67%, but it could be reached faster by vaccination, rather than natural immunization [30, 31].

Evidence for antibodies was found among participants who did not report having symptoms in the past eight months. Similar findings have been reported in many countries [19, 32] . These data underline the importance of testing asymptomatic individuals before travelling by air or other means of transport connecting different regions, even if they did not report any notion of exposure to SARS-CoV-2.

Our study has several limitations. Based on the assumption that 50% of the population are ≤18 years, we invited all residents from 50% of the households to participate in the study, while only people aged ≥18 were invited to participate in the remaining 50%. Participation was only 50% and A c c e p t e d M a n u s c r i p t 13 more women included which may limit the generalizability of our findings. Seroprevalence in age groups should be interpreted with caution as the age adjustment was done based on the 2019 population estimation [15] . Additionally, COVID-19-related symptoms that participants had developed in the previous eight months were reported retrospectively, resulting in a probable recall bias. We also probably missed some recent infections because we only tested the presence of IgG antibodies as illustrated by the lower sensitivity of our assay on a panel of samples collected between 1-30 days after symptom onset, suggesting a possible underestimation of recent infections.

The results of the first household SARS-CoV-2 serosurvey in Kinshasa show a high seroprevalence and spread in both the eastern and western regions of the city, illustrating that most cases were undiagnosed. These results provide an excellent picture of the extent of the COVID-19 pandemic in Kinshasa after the peak of the first wave, as well as lessons for adjusting the countermeasures. The country is now facing the second wave, which is apparently more contagious than the first one. Our findings therefore support strengthening of the testing capacity for both symptomatic and M a n u s c r i p t 21 

COVID-19: the current situation in the Democratic Republic of Congo

Public trust in the time of the coronavirus disease 2019 (Covid-19): the case of the DR Congo

United Nation High Commissioner for Refugees. Risk communication and community engagement (RCCE) -COVID-19

The impact of COVID-19 and strategies for mitigation and suppression in low and middle-income countries

COVID-19 on the African continent

The puzzle of the COVID-19 pandemic in Africa

Responding to the challenge of the dual Covid-19 and Ebola epidemics in the Democratic Republic of Congo-priorities for achieving control

Seroprevalence of anti-SARS-CoV-2 IgG antibodies in Kenyan blood donors

Prevalence of SARS-CoV-2 among high-risk populations in Lomé (Togo) in 2020

High SARS-CoV-2 Seroprevalence in healthcare workers in Bukavu, Eastern Democratic Republic of Congo

Challenges in interpreting SARS-CoV-2 serological results in African countries

Multiplex detection and dynamics of IgG antibodies to SARS-CoV2 and the highly pathogenic human coronaviruses SARS-CoV and MERS-CoV

The duration, dynamics and determinants of SARS-CoV-2 antibody responses in individual healthcare workers

The Democratic Republic of the Congo: World Development Indicators

findings from the second nationwide household serosurvey

Remarkable variability in SARS-CoV-2 antibodies across Brazilian regions: Nationwide serological household survey in 27 states

Seroprevalence of anti-SARS-CoV-2 IgG antibodies in Geneva, Switzerland (SEROCoV-POP): a population-based study

Prevalence of SARS-CoV-2 in six districts in Zambia in July, 2020 : a cross-sectional cluster sample survey

High SARS-CoV-2 seroprevalence in health care workers but relatively low numbers of deaths in urban Malawi

Seroprevalence of COVID-19 in Niger State

SARS-CoV-2 infection in Ivory Coast : a serosurveillance survey among gold mine workers

Seroprevalence of anti-SARS-CoV-2 IgG antibodies in Juba, South Sudan: a population-based study

Seroprevalence of SARS-CoV-2 in Guilan Province

Kinetics and seroprevalence of SARS-CoV-2 antibodies in children

Multi-organ impairment in low-risk individuals with long COVID

Immune determinants of COVID-19 disease presentation and severity

Population-based serosurvey for severe acute respiratory syndrome coronavirus 2 transmission

Potential antigenic cross-reactivity between severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and dengue Viruses

* Significant difference at 5%

A c c e p t e d M a n u s c r i p t