ABSTRACT
Population based data on COVID-19 are essential for guiding public policies. We report on the first of a series of planned seroprevalence surveys relying upon on household probabilistic samples of 133 large sentinel cities in Brazil, including 25,025 participants from all 26 states and the Federal District. Seroprevalence of antibodies to SARS-CoV-2, assessed using a lateral flow rapid test, varied markedly across the countrys cities and regions, from below 1% in most cities in the South and Center-West regions to up to 25% in the city of Breves in the Amazon (North) region. Eleven of the 15 cities with the highest seroprevalence were located in the North, including the six cities with highest prevalence which were located along a 2,000 km stretch of the Amazon river. Overall seroprevalence for the 90 cities with sample size of 200 or greater was 1.4% (95% CI 1.3-1.6). Extrapolating this figure to the population of these cities, which represent 25% of the countrys population, led to an estimate of 760,000 cases, as compared to the 104,782 cases reported in official statistics. Seroprevalence did not vary significantly between infancy and age 79 years, but fell by approximately two-thirds after age 80 years. Prevalence was highest among indigenous people (3.7%) and lowest among whites (0.6%), a difference which was maintained when analyses were restricted to the North region, where most indigenous people live. Our results suggest that pandemic is highly heterogenous, with rapid escalation in Brazils North and Northeast, and slow progression in the South and Center-West regions.
ABSTRACT
Population based data on COVID-19 are urgently needed for informing policy decisions, yet few such studies are available anywhere, as most surveys rely on self-selected volunteers. In the Brazilian State of Rio Grande do Sul (population 11.3 million), we are carrying out fortnightly household surveys in nine of the largest cities. Multi-stage probability sampling was used in each city to select 500 households, within which one resident was randomly chosen for testing. The Wondfo lateral flow rapid test for detecting antibodies against SARS-CoV-2 has been validated in four different settings, including our own, with pooled estimates of sensitivity (84.8%, 95% CI 81.4%;87.8%) and specificity (99.0%, 95% CI 97.8%;99.7%), which are within the acceptable range for epidemiological studies. In the first wave of the study (April 11-13), 4,188 subjects were tested, of whom two were positive (0.0477%; 95% confidence interval (CI) 0.0058%;0.1724%). In the second round (Apr 25-27) there were six positive subjects (0.1333%; 95% CI 0.0489%;0.2900%). We also tested family members of positive index cases, and nine out of 19 had positive results. Testing of reported COVID-19 cases according to RT-PCR confirmed that the test was highly sensitive under field conditions. The epidemic is at an early stage in the State, as the first case was reported on Feb 28, and by Apr 30, 50 deaths were registered. Strict lockdown measures were implemented in mid-March, and our results suggest that compliance was high, with full or near full compliance rates of 79.4% in the first and 71.7% in the second round. As far as we know, this is the only large population anywhere undergoing regular household serological surveys for COVID-19. The results show that the epidemic is at an early phase, and findings from the next rounds will allow us to document time trends and propose Public Health measures.
