key: cord-0722961-7rh8jkqn
authors: Etyang, Anthony O; Lucinde, Ruth; Karanja, Henry; Kalu, Catherine; Mugo, Daisy; Nyagwange, James; Gitonga, John; Tuju, James; Wanjiku, Perpetual; Karani, Angela; Mutua, Shadrack; Maroko, Hosea; Nzomo, Eddy; Maitha, Eric; Kamuri, Evanson; Kaugiria, Thuranira; Weru, Justus; Ochola, Lucy B; Kilimo, Nelson; Charo, Sande; Emukule, Namdala; Moracha, Wycliffe; Mukabi, David; Okuku, Rosemary; Ogutu, Monicah; Angujo, Barrack; Otiende, Mark; Bottomley, Christian; Otieno, Edward; Ndwiga, Leonard; Nyaguara, Amek; Voller, Shirine; Agoti, Charles; Nokes, David James; Ochola-Oyier, Lynette Isabella; Aman, Rashid; Amoth, Patrick; Mwangangi, Mercy; Kasera, Kadondi; Ng’ang’a, Wangari; Adetifa, Ifedayo; Kagucia, E Wangeci; Gallagher, Katherine; Uyoga, Sophie; Tsofa, Benjamin; Barasa, Edwine; Bejon, Philip; Scott, J Anthony G; Agweyu, Ambrose; Warimwe, George
title: Seroprevalence of Antibodies to SARS-CoV-2 among Health Care Workers in Kenya
date: 2021-04-24
journal: Clin Infect Dis
DOI: 10.1093/cid/ciab346
sha: 11ed251929fac01383e434752b1c0ad2f1e83c04
doc_id: 722961
cord_uid: 7rh8jkqn

BACKGROUND: Few studies have assessed the seroprevalence of antibodies against SARS-CoV-2 among Health Care Workers (HCWs) in Africa. We report findings from a survey among HCWs in three counties in Kenya. METHODS: We recruited 684 HCWs from Kilifi (rural), Busia (rural) and Nairobi (urban) counties. The serosurvey was conducted between 30th July 2020 and 4th December 2020. We tested for IgG antibodies to SARS-CoV-2 spike protein using ELISA. Assay sensitivity and specificity were 93% (95% CI 88-96%) and 99% (95% CI 98-99.5%), respectively. We adjusted prevalence estimates using Bayesian modeling to account for assay performance. RESULTS: Crude overall seroprevalence was 19.7% (135/684). After adjustment for assay performance seroprevalence was 20.8% (95% CrI 17.5-24.4%). Seroprevalence varied significantly (p<0.001) by site: 43.8% (CrI 35.8-52.2%) in Nairobi, 12.6% (CrI 8.8-17.1%) in Busia and 11.5% (CrI 7.2-17.6%) in Kilifi. In a multivariable model controlling for age, sex and site, professional cadre was not associated with differences in seroprevalence. CONCLUSION: These initial data demonstrate a high seroprevalence of antibodies to SARS-CoV-2 among HCWs in Kenya. There was significant variation in seroprevalence by region, but not by cadre.

Health care workers (HCWs) are critical in the acute-care response to epidemic waves of COVID-19, but they are also required to sustain normal health services beyond HCWs are considered to be at high risk of infection with SARS-CoV-2 1 . It is unclear whether the seroprevalence of SARS-CoV-2 antibodies among HCWs is more closely associated with community or hospital-based transmission risk as indicated by professional cadre. In some hospitals, seroprevalence was higher among cadres in lower paid jobs with little patient contact (e.g. housekeepers, porters) suggesting the source of infection may be their crowded living conditions rather than occupational risk 2 . The true extent of infection in HCWs in Kenya has been difficult to determine, due to factors including (1) the fact that a large proportion of infections (>90%) are asymptomatic 3 Serological surveys can estimate cumulative incidence of SARS-CoV-2 infection in either key groups, such as HCWs, or the general population 5 . They can also assess the effectiveness of infection prevention and control measures, which is important in sub-Saharan Africa (sSA)

where the availability of personal protective equipment and other preventive measures is constrained. To date HCW serosurveys in sSA have been limited to urban hospitals 6-8 ; there are no surveys from rural hospitals, where resources are even more constrained.

Serosurveys on different population groups or in different geographical regions can also inform vaccine prioritization policies. This is especially important in LMICs where only a small proportion of the population are likely to receive vaccines in the early phase of the vaccine campaign 9 .

A c c e p t e d M a n u s c r i p t 5 Because the presence of antibodies to SARS-CoV-2 appears to be strongly protective against repeat infection over a 6-month period 10, 11 , knowledge of past infection could be useful for avoiding unnecessary quarantines which would help preserve the limited numbers of personnel available to deal with the pandemic and other health needs in the region.

We report initial findings from SARS-CoV-2 antibody testing from HCWs in three sites in Coastal, Central, and Western Kenya.

Study sites ( Figure S1 ) were selected after consultation with the individual county COVID-19

Rapid Response Teams (RRTs). For Kilifi County, a predominantly rural area located on the Indian Ocean coast, we enrolled participants at Kilifi County Hospital, which is the main referral facility in the region. For Busia County, which is also predominantly rural and located in the western region of Kenya, we enrolled HCWs at Busia County Referral Hospital, the main referral facility in the area, and two other facilities in the county; Alupe Sub-County Hospital which has been designated as the isolation facility for COVID-19 patients in the county, and Kocholia Sub-County Hospital. In Nairobi County, the capital city of Kenya, we enrolled HCWs at the Kenyatta National Hospital (KNH), the main referral facility for the city as well as the country 12 .

We used a variety of strategies to recruit a convenience sample of HCWs at each of the study sites, including word of mouth, advertising at hospital notice boards and messages sent via mobile phone. HCWs of all cadres were eligible to participate in the study. In Kilifi and Busia we aimed to recruit ≥ 50% (N=441) of the 882 HCWs working in the healthcare facilities, which we considered to be both feasible and likely to provide a representative A c c e p t e d M a n u s c r i p t 6 sample. We used a slightly different strategy at KNH, where the primary aim of the study was to determine incidence and antibody kinetics among ~180 HCWs, comprising ~4% of the hospital's estimated 5,000 HCWs 12 l, who were likely (by self-report) to be available for a

year-long longitudinal study.

Serosurveillance was conducted as a public health activity requested by the Kenya Ministry of Health and ethical approval for collection and publication of these data was obtained from the Kenya Medical Research Institute Scientific and Ethics Review Unit (KEMRI/SERU/CGMR-C/203/4085). HCWs provided written and/or verbal informed consent for participation in the study. Results of the antibody testing were reported confidentially to each HCW together with information explaining the implications of the test results.

The study took place between 30 th July 2020 and 4 th December 2020. Data collection was performed by members of staff from the participating hospitals, trained on the study procedures.

We collected 6ml of venous blood in sodium heparin tubes from each participant. Serum was obtained by centrifuging the samples at 450 x g for 5 minutes before storage at -80ºC.

Samples were then transported in dry ice to the KEMRI-Wellcome Trust research laboratories in Kilifi for assays.

A simple one-page questionnaire (provided in the appendix) was administered to the HCWs either electronically or on paper, in which data on demographic and clinical characteristics were collected.

A c c e p t e d M a n u s c r i p t 7

All samples were tested at the KWTRP laboratories in Kilifi for IgG to SARS-CoV-2 whole spike protein using an adaptation of the Krammer Enzyme Linked Immunosorbent Assay (ELISA) 13 . We assumed an assay sensitivity of 92.7% (95% CI 87.9-96.1%) and specificity of 99.0% (95% CI 98.1-99.5) based on validation studies that we had previously conducted 14 .

Results were expressed as the ratio of test OD to the OD of the plate negative control; samples with OD ratios greater than two were considered positive for SARS-CoV-2 IgG.

Continuous variables were summarized as means and standard deviations if normally distributed and medians with interquartile ranges for non-normally distributed variables.

Categorical data were presented as counts and percentages. Bayesian modelling was used to adjust seroprevalence estimates for the sensitivity and specificity of the assay. Noninformative priors were used for all parameters, and the models were fitted using the Rstan software package 15 (see appendix for code). We tested for associations between seroprevalence and professional cadre and site, respectively using multivariable logistic regression.

All analyses were conducted using Stata™ Version 15 software (College Station, Texas, USA) and R version 3.6.1 (Vienna, Austria).

We recruited 684 HCWs from Nairobi, Busia and Kilifi ( Figure S2 and Table 1 ). The numbers of the HCWs that we recruited as a proportion of total number of staff at the facilities were 70% in Kilifi, 50% in Busia, and ~4% in Nairobi. The mean age ± SD of the participants was 35 A c c e p t e d M a n u s c r i p t 8 ± 11 years and 54% were female. Sixteen (2%) of the HCWs reported that they had acute respiratory symptoms at the time of sample collection.

Out of the 684 HCWs, 135 (19.7%) were seropositive for antibodies to SARS-CoV-2 (Table 2) .

After adjusting for test performance characteristics, the seroprevalence was 20.8% (95% Credible 1-0.4) were less likely to be seropositive compared to those in Nairobi. Professional cadre, age and sex were not associated with seroprevalence in both univariable and multivariable analyses. Site-specific analyses also did not reveal any association between seroprevalence and professional cadre (Table S1 ).

We report results of a SARS-CoV-2 seroprevalence study conducted among HCWs in 3 counties in Kenya. We found an overall seroprevalence of SARS-CoV-2 antibodies of 20.8% (95% CrI 17.5-24.4%). There were significant differences in seroprevalence associated with hospital region, but no differences associated with professional cadre.

Our estimates of seroprevalence are higher than what was found in most of studies from Africa that have been published to date, all of which were conducted in urban areas 6-8, 16 and A c c e p t e d M a n u s c r i p t 9 had a pooled seroprevalence of 8.2% (95% CI 0. 8-22.3) 17 . We conducted our study during and shortly after the first wave of the epidemic in Kenya 4 , while the previous studies in Africa were conducted relatively early in the epidemic. Our estimates are similar to those observed among HCWs in several high-income countries at the peak of their first wave of the epidemic 17 .

Consistent with other studies conducted in Kenya 4, 14, [18] [19] [20] , we found significant differences in seroprevalence by region. HCWs in urban Nairobi had significantly higher seroprevalence than those in Busia and Kilifi, which are rural counties. Studies in Spain and India have also shown significant regional differences, with higher seroprevalence in urban areas, such as Madrid and New Delhi, compared to rural areas 21, 22 . However, even in the rural counties in Kenya, HCWs had seroprevalence estimates that were similar to those in HCWs in urban areas in Spain 23 , USA 24 and Malawi 6 .

We found no differences in seroprevalence by professional category even when the analyses were stratified by study site. The absence of differences in seroprevalence by cadre in the presence of significant differences by geographical region suggests that community transmission could be playing a bigger role than workplace exposure. In studies of HCWs conducted in the UK, the incidence of infection mirrored that seen in the community 2, 25 . This suggests that efforts to suppress community transmission are likely to reduce infections among HCWs.

The results of this study provide further evidence that there has been significant undocumented transmission of the SARS-CoV-2 virus within Kenya. Additional evidence of significant undocumented transmission in Kenya derives from (1) two studies of seroprevalence among blood transfusion donors 14, 18 ; (2) a study of truck drivers and their assistants conducted at the same time as this survey in Kilifi and Busia that found a seroprevalence of 42% 19 , and; (3) In a study of antenatal clinic attendees, seroprevalence A c c e p t e d M a n u s c r i p t 10 was 50% at Kenyatta National Hospital in August 2020, and 11% at Kilifi County Hospital in November 2020 20 .

A particular strength of this study is that we conducted it in several sites, which enabled us to detect a significant burden of infection among HCWs in rural parts of the country.

Another strength is that we used an assay that was validated using both local and external samples and which performed well in a WHO-sponsored international standardization study 26 . Although we adjusted our figures using Bayesian modelling to take into account assay performance, the reported seroprevalence could still be underestimated due to antibody waning 27 . The longitudinal phase of the current study will help address this issue.

Another possible reason for underestimation of the prevalence in our study would be spectrum bias 28 since the samples that we used in validating the assay, although derived from the local population, these individuals were not necessarily the same as the HCWs that participated in the present survey.

Our study had several limitations. We did not perform genetic sequencing to establish the likely sources of infections among the HCWs, although as argued above, the data we obtained suggests that community transmission was the main driver of infections among the HCWs. The non-random selection of only a small proportion of the HCWs in Nairobi could have led to an overestimation of the seroprevalence if the HCWs sampled had an overrepresentation of individuals who had experienced symptoms in the past. However, this would have also resulted in a higher proportion of HCWs in Nairobi having positive results from previously conducted PCR tests, but we did not observe this. In addition a household survey found that 35% of the population in Nairobi had antibodies to SARS-CoV-2 29 , and the rural-urban difference in seroprevalence among HCWs that we observed was similar to what has been observed in other studies conducted in Kenya 14, [18] [19] [20] M a n u s c r i p t 11

In conclusion, we found a high prevalence of antibodies to SARS-CoV-2 among HCWs in Kenya, with significant regional differences and no differences based on cadre. The results suggest that infection with SARS-CoV-2 among HCWs is driven more by background population levels of infection than workplace exposure and will be useful in informing measures to control the on-going pandemic.

A c c e p t e d M a n u s c r i p t 12

We would like to thank all HCWs who participated in the study as well as the county health teams and the Kenya Paediatric Research Consortium (KEPRECON) that facilitated data collection for the study. We thank F. Krammer for providing the plasmids used to generate the RBD, spike protein, and CR3022 monoclonal antibody used in this work. Development of SARS-CoV-2 reagents was partially supported by the NIAID Centres of Excellence for 

Epidemiology of and Risk Factors for Coronavirus Infection in Health Care Workers: A Living Rapid Review

SARS-CoV-2 seroprevalence and asymptomatic viral carriage in healthcare workers: a cross-sectional study

Kenyan Ministry of Health: COVID-19 situation reports

Revealing the extent of the COVID-19 pandemic in Kenya based on serological and PCR-test data

The Power of Antibody-Based Surveillance

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

SARS-CoV-2 Seropositivity in Asymptomatic Frontline Health Workers in Ibadan, Nigeria

High SARS-CoV-2 Seroprevalence in Healthcare Workers in

Eastern Democratic Republic of Congo

Ensuring global access to COVID-19 vaccines

University Hospitals Staff Testing G. Antibody Status and Incidence of SARS-CoV-2 Infection in Health Care Workers

Do antibody positive healthcare workers have lower SARS-CoV-2 infection rates than antibody negative healthcare workers? Large multi-centre prospective cohort study (the SIREN study)

A serological assay to detect SARS-CoV-2 seroconversion in humans

IgG antibodies in Kenyan blood donors

RStan: the R interface to Stan

Seroprevalence of antibodies against SARS-CoV-2 among public community and health-care workers in Alzintan City of Libya

Seroprevalence of SARS-CoV-2 antibodies and associated factors in health care workers: a systematic review and meta-analysis

Temporal trends of SARS-CoV-2 seroprevalence in transfusion blood donors during the first wave of the COVID-19 epidemic in Kenya

Seroprevalence of anti-SARS-CoV-2 IgG antibodies among truck drivers and assistants in Kenya

Sero-surveillance for IgG to SARS-CoV-2 at antenatal care clinics in two Kenyan referral hospitals

ENE-COVID): a nationwide, population-based seroepidemiological study

findings from the second nationwide household serosurvey

Seroprevalence of antibodies against SARS-CoV-2 among health care workers in a large Spanish reference hospital

Prevalence of SARS-CoV-2 Antibodies in Health Care Personnel in the

COVID-19: PCR screening of asymptomatic health-care workers at London hospital. The Lancet

Page and the collaborative study participants Establishment of the WHO International Standard and Reference Panel for anti-SARS-CoV-2 antibody

Declining prevalence of antibody positivity to SARS-CoV-2: a community study of 365,000 adults

Are Seroprevalence Estimates for Severe Acute Respiratory Syndrome Coronavirus 2 Biased?

Antibody Survey in Nairobi City County Conducted in

A c c e p t e d M a n u s c r i p t 14 A c c e p t e d M a n u s c r i p t 19 3 (11.3-51.9) Laboratory 64 12 18.8 20.4 (10.6-31.5) Other 168 28 16.7 17.5 (11.8-24. 3)

Kilifi 200 23 11.5 11.9 (7.2-17.6) Nairobi 183