Study population and setting
The study objective was to determine the burden of COVID-19 infection in a rural South American setting to assess whether the 266% increase in mortality from April and May 2020 was related to increased COVID-19 infection in the area. The study was conducted in the village of Atahualpa, Ecuador, on the coast. The village has electricity and piped water, though most houses use an open-latrine system. A parent study, the Atahualpa Project, enrolled 730 individuals in the village who were 40 or more years of age since 2012, and who were actively enrolled as of 2019, as a prospective population-based cohort with annual visits. 673 participants from the Atahualpa Project were included in their analysis. They collected data on COVID-19 symptoms in the past 2 months, and used an IgM and IgG lateral flow antibody test from finger prick blood to determine seroprevalence. They also included covariates around age, gender, highest level of education attained, size of household and household characteristics, whether they were confined in the home, and other cardiovascular health metrics.
Summary of Main Findings
Of the 730 participants of 2019, 57 died, emigrated, or refused to participate in the sub-study. Of the 673, 303 (45.0%) were seropositive for SARS-CoV-2. The majority tested positive for both positive IgM and IgG, 9 (3.0% of 303), with only positive IgM indicating a current or recent infection, and 47 (15.5% of 303) with only positive IgG suggesting a past infection. Of the seropositive, 77% (n=232) reported COVID-19 symptoms, while 14.9% of seronegative people (n=55/370) reported COVID-19 symptoms. Seropositive individuals reported exposure to open latrines and less home confinement in the past 2 months compared to seronegative persons. After adjustment for all covariates, using open latrines increased the odds of being seropositive by more than two-fold (OR: 2.14, 95% CI: 1.23 – 1.72).
The study describes the prevalence of SARS-CoV-2 antibodies among a rural population not often described in literature. Of the 730 enrolled in the parent study, there was a high rate of response (92.2%), which suggests this substudy likely has a representative sample of the parent study. They also collected a number of variables that are not often included or considered in high-resource settings, such as the use of latrines, which leads to important novel findings. The use of an antibody test, as well as assessing symptoms, allows them to identify potential asymptomatic cases in this population.
By using an antibody test, they are likely missing individuals who recently became infected; they also would miss any adults that were infected but did not develop an antibody response. Additionally, while the specific antibody test they used had a high sensitivity/specificity compared to an ELISA assay, their reference standard is imperfect, and has its own sensitivity/specificity limitations. Therefore, there may be some misclassification of cases/controls based on the test and the subsequent qualitative “positive/negative” results as interpreted by the study team. Seroprevalence is likely still very high, but it is possible there could be some overestimation related to this interpretation. The study also only included individuals at least 40 years of age, which may undercount or overcount the degree of spread within the village if those less than 40 had a lower or higher infection rate than other age strata. By asking about symptoms in the prior two months, there may be recall issues over the whole period, with some misclassification. Finally, they only report symptoms in the aggregate, and were unable to assess whether any particular symptoms were associated with seropositivity in this population.
This is one of the first studies to estimate the burden of SARS-CoV-2 infection in a rural village setting.
This review was posted on: 3 September 2020