ABSTRACT
In sub-Saharan Africa, reported COVID-19 numbers have been lower than anticipated, even when considering populations younger age. The extent to which risk factors, established in industrialised countries, impact the risk of infection and of disease in populations in sub-Saharan Africa, remains unclear. We estimated the incidence of mild and moderate COVID-19 in urban Mozambique and analysed factors associated with infection and disease in a population-based surveillance study. During December 2020-March 2022, households of a population cohort in Polana Canico, Maputo, Mozambique, were contacted biweekly. Residents reporting any respiratory sign, anosmia, or ageusia, were asked to self-administer a nasal swab, for SARS-CoV-2 PCR testing. Of a subset of 1400 participants, dried blood spots were repeatedly collected three-monthly from finger pricks at home. Antibodies against SARS-CoV-2 spike glycoprotein and nucleocapsid protein were detected using an in-house developed multiplex antibody assay. We estimated the incidence of respiratory illness and COVID-19, and SARS-CoV-2 seroprevalence. We used Cox regression models, adjusting for age and sex, to identify factors associated with first symptomatic COVID-19 and with SARS-CoV-2 sero-conversion in the first six months. During 11925 household visits in 1561 households, covering 6049 participants (median 21 years, 54.8% female, 7.3% disclosed HIV positive), 1895.9 person-years were followed up. Per 1000 person-years, 364.5 (95%CI 352.8-376.1) respiratory illness episodes of which 72.2 (95%CI 60.6-83.9) COVID-19 confirmed, were reported. Of 1412 participants, 2185 blood samples were tested (median 30.6 years, 55.2% female). Sero-prevalence rose from 4.8% (95%CI 1.1-8.6%) in December 2020 to 34.7% (95%CI 20.2-49.3%) in June 2021, when 3.0% were vaccinated. Increasing age (strong gradient in hazard ratio, HR, up to 15.70 in [≥]70 year olds, 95%CI 3.74-65.97), leukaemia, chronic lung disease, hypertension, and overweight increased risk of COVID-19. We found no increased risk of COVID-19 in people with HIV or tuberculosis. Risk of COVID-19 was lower among residents in the lowest socio-economic quintile (HR 0.16, 95%CI 0.04-0.64), with no or limited handwashing facilities, and who shared bedrooms (HR 0.42, 95%CI 0.25-0.72). Older age also increased the risk of SARS-CoV-2 seroconversion (HR 1.57 in 60-69 year olds, 95%CI 1.03-2.39). We found no associations between SARS-CoV-2 infection risk and socio-economic, behavioural factors and comorbidities. Active surveillance in an urban population cohort confirmed frequent COVID-19 underreporting, yet indicated that the large majority of cases were mild and non-febrile. In contrast to industrialised countries, deprivation did not increase the risk of infection nor disease.