Just 9.3% of U.S. adults were seropositive for COVID-19 as of July (95% CI 8.8%-9.9%), based on a nationwide sample of dialysis patients extrapolated to the U.S. population, researchers found.
But the figure was substantially higher in the northeastern U.S., the nation’s hardest hit region, where an estimated 27.2% of the national dialysis population was seropositive (95% CI 25.9%-28.5%), reported Shuchi Anand, MD, of Stanford University in Palo Alto, California, and colleagues in The Lancet. In western states, estimated seroprevalence was 3.5%.
Anand and colleagues also found relatively few of the seropositives had actually been diagnosed with COVID-19: just 9.2% (95% CI 8.7%-9.8%), via comparison of seroprevalence and case counts per 100,000 population as of June 15.
Co-author Julie Parsonnet, MD, also of Stanford University, said in a statement that the study “clearly confirms that despite high rates of COVID-19 in the United States, the number of people with antibodies is still low and we haven’t come close to achieving herd immunity” — which normally requires around 70% of a population to have been exposed and become immune.
An accompanying editorial by Barnaby Flower, PhD, and Christina Atchison, PhD, both of Imperial College London in England, pointed out some advantages for the new study over prior seroprevalence surveys. First, earlier surveys used “sampling strategies prone to selection bias,” which is problematic for a virus that “disproportionately affects some ethnic groups and deprived communities who are less likely to participate in research.” Second, the large surveys used point-of-care lateral flow assays, which are less sensitive than the laboratory assays performed by Anand’s group.
“This adds uncertainty and necessitates substantial adjustment of raw data to account for false-negative results,” they wrote.
Flower and Atchison praised the Siemens chemiluminescence assay used in the current study, noting it was “the best-performing platform in the largest external appraisal of commercial assays to date, in terms of both sensitivity and specificity.”
Anand and colleagues hypothesized that patients receiving dialysis might be “an ideal sentinel population in which to study the evolution of the COVID-19 public health crisis,” since not only do they routinely undergo laboratory studies, but the U.S. dialysis population is similar to those at risk of severe COVID-19, with “advanced age, non-white race, poverty and diabetes” as the rule rather than the exception in this population.
“We were able to determine – with a high level of precision – differences in seroprevalence among patient groups within and across regions of the United States, providing a very rich picture of the first wave of the COVID-19 outbreak that can hopefully help inform strategies to curb the epidemic moving forward by targeting vulnerable populations,” Anand said in a statement.
For the study, Anand and colleagues partnered with a nationwide network of around 1,300 dialysis facilities that serve approximately 65,000 patients. Other inputs included patient-level residence data with cumulative daily cases and deaths per 100,000 by Johns Hopkins University and nasal swab test positivity rates at a state level by the Covid Tracking Project. Residency information was also linked with zip code-level socioeconomic data from the 2018 American Community Survey’s 5-year estimates.
Overall, 28,503 patients were included from 46 states with similar age, sex, race and ethnicity distribution to the U.S. dialysis population. As expected, the dialysis population had more older people, men, and people living in majority Black and Hispanic neighborhoods than in the general population.
Residents of predominantly non-Hispanic Black (OR 3.9, 95% CI 3.4-4.6) and Hispanic (OR 2.3, 95% CI 1.9-2.6) neighborhoods were more likely to be seropositive than those living in non-Hispanic white neighborhoods. Highest population density also was tied to increased seropositivity rates.
But Flower and Atchison also noted substantial limitations to the study: for one, extrapolating seroprevalence in dialysis patients to the general population is “inherently problematic,” even after adjustments. This population attends a health facility three times a week, meaning their risk of SARS-CoV-2 is likely higher. Patients with end stage kidney disease are also less likely to mount a detectable antibody response and more likely to die from COVID-19.
Nevertheless, the editorialists applauded Anand’s group for this “blueprint” of a scalable sampling strategy for national surveillance in the dialysis population.
“Although general population estimates from dialysis sampling are imperfect, they at least remain consistent across the country and from one survey to the next, permitting longitudinal surveillance,” Flower and Atchison wrote.
This study was supported by Ascend Clinical Laboratories.
Anand disclosed support from the NIH. Other co-authors disclosed support from the NIH. Three co-authors disclosed employment by Ascend Clinical Laboratories. One co-author disclosed being on the Board of Directors of Satellite Healthcare, a not-for-profit dialysis organization.
Flower and Atchison disclosed no conflicts of interest.