Skip to main content

Seroprevalence of Antibodies to SARS-CoV-2 in 10 Sites in the United States, March 23-May 12, 2020

Our take —

This cross-sectional study aimed to estimate the prevalence of SAR-Cov-2 infection through antibody detection across 10 cities in the United States. Age and sex-standardized seroprevalence estimates using weights derived from the US census were used to estimate prevalence. The estimated proportion of the site populations with detectable SARS-Cov-2 antibodies ranged from 1.0% to 6.9%. The estimated number of infections was much greater than the number of reported cases in all sites, with estimates 6 to 24 times higher than those cases reported by health departments. This study suggests a large underestimation of SARS-Cov-2 infections by surveillance methods, though given samples were ascertained using routine clinical samples rather than random households sampling, caution is warranted and results may be overestimated if patients were seeking care for COVID-19 symptoms or underestimated if those engaged in care during the pandemic were more likely to have co-morbidities which result in behavioral change to minimize exposure. Though estimates suggest several asymptomatic or otherwise uncounted infections, most in the population remain susceptible.

Study design

Cross-Sectional

Study population and setting

This cross-sectional study aimed to estimate the prevalence of SAR-Cov-2 infection through antibody detection. Residual sera was used from routine clinical testing by two laboratory companies from March 23 to May 12, 2020 in the San Francisco Bay area, California; Connecticut; south Florida; Louisiana; Minneapolis-St Paul-St Cloud metro area, Minnesota; Missouri; New York City metro area, New York; Philadelphia metro area, Pennsylvania; Utah; and western Washington State. This convenience sample of 16,025 individuals was inclusive of all age groups. Antibodies to SARS-CoV-2 spike protein were measured using an enzyme-linked immunosorbent assay. Age-standardized and sex-standardized seroprevalence estimates using weights derived from US census were used to estimate prevalence within 10 sites.

Summary of Main Findings

The estimated proportion of the site populations with detectable SARS-Cov-2 antibodies ranged from 1.0% (San Francisco Bay area) to 6.9% (New York City. Compared to cases reported by health departments, the estimated proportion of the population infected was 6 to 24 times higher across the 10 sites.

Study Strengths

This study leveraged a large sample (n=16,025) of routine clinical data, and standardized seroprevalence estimated by age and by sex.

Limitations

Samples were obtained from routine clinical care, and therefore may not represent the broader population. Especially given changes in uptake and provision of healthcare during the outbreak, cases may represent those who may have more underlying or severe health conditions or may include individuals seeking care related to COVID-19. It is possible that the ELISA may exhibit cross-reactivity with antibodies to other common human coronaviruses and therefore false-positive results for SARS-CoV-2 maybe present in this sample.

Value added

This study leveraged a large sample of routine clinical data to estimate the proportion of the population with evidence of previous infection with SARSCoV-2 in 10 geographically diverse US sites with known community transmission.

This review was posted on: 17 August 2020