key: cord-0744448-7qk4abxl
authors: Naranbhai, Vivek; Chang, Christina C; Beltran, Wilfredo F Garcia; Miller, Tyler E; Astudillo, Michael G; Villalba, Julian A; Yang, Diane; Gelfand, Jeffrey; Bernstein, Bradley E; Feldman, Jared; Hauser, Blake M; Caradonna, Timothy M; Alter, Galit; Murali, Mandakolathur R; Jasrasaria, Rashmi; Quinlan, Joan; Xerras, Dean C; Betancourt, Joseph R; Louis, David N; Schmidt, Aaron G; Lennerz, Jochen; Poznansky, Mark C; Iafrate, A John
title: High seroprevalence of anti-SARS-CoV-2 antibodies in Chelsea, Massachusetts
date: 2020-09-09
journal: J Infect Dis
DOI: 10.1093/infdis/jiaa579
sha: 031f37b4693fe260206c07b6de4337f73ce75571
doc_id: 744448
cord_uid: 7qk4abxl

SARS-CoV-2 antibody testing allows quantitative determination of disease prevalence, especially important in high-risk communities. We performed anonymized convenience sampling of 200 currently asymptomatic residents of Chelsea, the epicenter of COVID-19 illness in Massachusetts by BioMedomics SARS-CoV-2 combined IgM-IgG point-of-care lateral flow immunoassay. The seroprevalence was 31.5% (17.5% IgM+IgG+, 9.0% IgM+IgG- and 5.0% IgM-IgG+). 50.5% of participants reported no symptoms in the preceding 4 weeks, of which 24.8% (25/101) were seropositive, and 60% of these were IgM+IgG-. These data are the highest seroprevalence rates observed to date and highlight the significant burden of asymptomatic infection.

The diagnosis of COVID-19 illness is based on clinical symptoms and detection of SARS-CoV-2 by polymerase chain reaction (PCR), and serologic testing may aid in diagnosis and in estimating disease prevalence. We have validated the BioMedomics SARS-CoV-2 IgM/IgG LFA point-of-care (POC) lateral flow immunoassay (LFA) as a Laboratory Developed Test (LDT) in a high-complexity lab. This assay identifies anti-receptor binding domain (RBD) IgM and IgG.

Motivated by a clinical observation that COVID-19 inpatients were enriched for residents from the City of Chelsea, and public data showing Chelsea had the highest cumulative COVID-19 case rate (1890 per 100,000 persons; 712 cases) on 14 April 2020 (1), we performed a rapid, pilot, seroprevalence study at a mobile testing site in Chelsea.

Over two consecutive afternoons (April [14] [15] 2020) , at a mobile testing site at Bellingham Square, a central square in the City of Chelsea that abuts a bus commuter junction,. we enrolled 200 interested consenting participants who were Chelsea residents, aged ≥18 years, with no current symptoms and no history of a positive SARS-CoV-2 PCR test. We obtained verbal consent for an anonymized questionnaire and serological testing with no return of results. To minimize ascertainment bias we did not have any prior advertisement for the study and did not actively recruit individuals locally or online. An information poster in English, Spanish and Portuguese was available at our site (Supplementary Figure 1) . Prospective participants were given a surgical mask and directed to a spaced queue to await discussion with a study investigators and given a copy of the information poster. We had substantial interest within minutes of commencing testing and estimate an average queue length of 7 individuals and a <5% drop-out rate. We did not systematically A c c e p t e d M a n u s c r i p t document the refusal/decline frequency nor the fraction of individuals on day two who attended through referral as this was not feasible.

Participants were provided advice on precautions, hand sanitizer or soap and face masks, and were compensated with a USD $5 voucher. The study was performed with approval of the Partners Institutional review board (#2020P001081) and the city manager.

The brief COVID-19-focused anonymized questionnaire was available in three languages (English, Spanish and Portuguese, Supplementary Figure 2 ) and administered in the participant's preferred language by two trilingual doctors.

Participants were tested with the BioMedomics SARS-CoV-2 combined IgM/IgG LFA (BioMedomics, Morrisville, NC), according to the manufacturer's recommendations (2) . Twenty microliters of blood were obtained using a fingerstick lancet (BD Microtainer lancet, Franklin Lakes, NJ), and applied immediately to the device. This was read after 10 minutes by one of two trained doctors. Positive, weak positive and negative bands for control, IgM and IgG were recorded and a photograph obtained. A second reader reviewed the photographs blinded to the field-results and agreement calculated, and consensus reached on discrepant readings. Our inability to return results was reiterated prior to the fingerstick.

To assess LFA cross-reactivity we used an ELISA in which purified receptor binding domain (RBD) proteins from two common-cold coronaviruses (HKU1 and NL63) and SARS-CoV-2 (SARS2) were coated on ELISA plates. Detailed methods have been presented elsewhere (3) .

A c c e p t e d M a n u s c r i p t

Data were entered and analysed by R (v4.0, the R Foundation). Descriptive summary statistics and regression models were used for the overall group and to compare those who were seropositive to those who were seronegative. We adjusted estimates of prevalence for the sensitivity and specificity of the assay per methods described by Larremore et al. (4) , using an online tool: https://larremorelab.github.io/covid-calculator1.

Over two consecutive afternoons totaling ~9 hours, we recruited 200 currently asymptomatic Chelsea residents, who had not previously tested positive for COVID-19, to participate in this anonymous study.

The median age was 46 years (interquartile range (IQR) 27-55) and 40% were female ( Table 1 ). The median number of cohabiting adults was 2 and the median number of cohabiting children was 1.

Nearly half, 42.4% (84/198), had to continue to leave home for work during the COVID-19 epidemic and nearly a quarter 23.6% (47/199) reported a known COVID-19 contact.

While all reported no current symptoms, 99 (48.5%) participants reported COVID-19-like symptoms in the preceding 4 weeks, occurring a median of 11 (IQR 5-14) days prior. The most common symptoms reported were cough (26.5%), rhinitis (24.0%), sore throat (23.5%) and myalgia (23.0%) ( A c c e p t e d M a n u s c r i p t

We validated the Biomedomics LFA assay using blood samples from 57 inpatients admitted for COVID-19 with PCR confirmed SARS-CoV-2 infection and observed an overall sensitivity of 80% for IgM, 85% for IgG, and 90% for combined IgM or IgG more than 14 days after symptom onset. The fraction testing seropositive increased steadily over these first 14 days. Assay specificity was determined to be >99% based on testing of 263 pre-COVID-19 outbreak specimens and 114 asymptomatic blood donors during the early outbreak (Table 1) . We did not identify evidence of cross-reactivity (Supplementary Figure 3) .

The overall seroprevalence (IgM+ and/ or IgG+) was 31.5% (63/200) ( Table 2 ). In detail, 17.5% were positive for both IgM and IgG (IgM+IgG+), 9.0% were IgM+IgG-, and 5.0% were IgM-IgG+. Overall, 26.5% of the cohort were IgM+ and 14.0% IgG+. Inter-reader agreement was 97%. Adjusting for the sensitivity and specificity of the assay, the estimated prevalence was 32.7% (90% Credible interval 26.4%-39.4%)for IgM, 16% (90% credible interval 11.4%-21.0%) for IgG and32.8% (27.2%-38.8%) for combined IgM or IgG.

Seropositive participants had a female predominance, a higher median number of cohabiting children, symptomatology in the last 4 weeks, with a more distant onset of symptoms (Table 1) . A reduced sense of smell or taste was reported in 28.6% vs. 5.8% in the seropositive and seronegative participants respectively.

In a multivariable regression model, the number of cohabiting children was an independent risk factor for seropositivity (Supplementary Table 1 

This study performed in a convenience sample of currently asymptomatic adults at a busy commuter junction with no previous diagnosis of COVID-19 has the highest reported seroprevalence of SARS-CoV-2 antibodies to date at 31.5%. This is 16.7-fold the 1.89% case rate based on symptomatic PCRbased testing at that time. About half were asymptomatic for the last 4 weeks. These data indicate significant community transmission and asymptomatic infection notwithstanding that this was not a population representative sample.

Massachusetts has the fifth highest number of COVID-19 cases in the USA (5) and has since recorded 100,000 confirmed cases (6). Within MA, Chelsea remains the city with the highest cumulative case rate, at 7.23% (May 27 2020) (7) . Notably, 40.2% of 6,742 tests performed were PCR-positive (7) .

The city of Chelsea is 4 miles from the city of Boston, and is the smallest city in MA at 2.21 square A c c e p t e d M a n u s c r i p t miles (sqm) and a high population density at 17959.28/sqm in comparison to that of Massachusetts (883.64/sqm) or Boston (16381.27/sqm) (8) . Based on population estimates from 2019, 66.9% were Hispanic or Latino, 45.5% were foreign-born and 70.3% spoke a language other than English at home (8) . The per capita income was $24,338, 18.8% persons were in poverty and 17.5% held a bachelor's degree and above (8) . Our results suggest that, at least amongst the tested population, many individuals may not be able to effectively socially distancedue to high population density, continued work-attendance as essential workers, poor access to care and other socioeconomic barriers.

after IgM wanes. This study corroborates recent reports of serological testing in SARS-CoV-2 that challenge this dogma, with early detection of IgG (as early as day 4 after symptom onset and only 1 day later than IgM) and occasionally detectable IgG prior to IgM (9) . A recent, preprint integrated data from 22 studies finding significant variation in seroconversion with a mean of 12.6 and 13.3 days (standard deviation 5.7-5.8 days) post symptom onset for IgM and IgG, and notably both IgM and IgG may be detected as early as day 0 (10). IgM titer peaked at around day 25 and decreased significantly shortly thereafter, whereas IgG peaked around day 25-27 and remained high for the reportable duration (up to 60 days) (10) . We noted that persons recently symptomatic already developed IgG, and an IgM+IgG-response predominated in the 4-week-asymptomatic group. While the latter may reflect early infection, it may also be plausible that their viral burdens were low perhaps affecting kinetics of class-switching. If IgG is durable and protective against future reinfections, these asymptomatic or pauci-symptomatic individuals may remain susceptible. None had been previously tested, reflecting missed opportunities in contact-tracing strategies.

Several US population-based COVID-19 seroprevalence studies were conducted in March-April 2020 with varying methodologies, patient populations and contrasting results. The Santa Clara study tested 1,702 residents with the Premier Biotech IgM-IgG POC LFA, of whom 4.65% were seropositive (12) . Using the Abbott Architect IgG test, seroprevalence was 1.79% in 4,865 Boise, Idaho residents (13) and 0.1% in 1,000 blood-donors in San Francisco Bay (14) ; In contrast, this was 12.5% in 15,101 New York State and 22.7% in New York City residents based on dry-blood spots (15) .

This mobile pilot study was limited by its sample size and random sampling; it was not designed to represent population structure. At the time of this study, there were no issued management guidelines for seropositive individuals. Further, sensitivities around undocumented migrants compelled anonymized testing. Thus, we could not deliver, confirm nor follow these results. We excluded children from this pilot but our data suggest asymptomatic infection in children may play a role in transmission dynamics and deserves further study. We were careful to reiterate that serological tests were not yet approved by the US FDA and remain a research tool. We emphasized that antibody detection does not equate antibody function and the questions of reinfection and seroprotection remain unanswered. Acknowledging all the unknowns of herd immunity in this new disease, 31.5% though high is still far from the estimated seroprevalence consistent with herd protection.

Our study demonstrated a remarkably high seroprevalence of 31.5% in a vulnerable urban population despite being currently asymptomatic. Serologic testing can aid in understanding community prevalence, and uncover the substantial burdens of missed, mild or asymptomatic infections particularly in settings where PCR-tests are limited. Enhanced testing, contact tracing, social-distancing and isolation are particularly needed in vulnerable communities. M a n u s c r i p t

We would like to acknowledge the participants of the study. We thank the manager of the City of Chelsea, Tom Ambrosino, Mimi Graney and the city volunteers and security personnel provided during the conduct of study.

Footnote page:

All authors: none declared 

The results of this study was presented in part, at a local meeting involving the City of Chelsea and the leadership of Partners Healthcare via Zoom meeting on 17 April, 2020 and also summarized via video for the community via the Chelsea Community Cable on 20 April, 2020. M a n u s c r i p t 

A c c e p t e d M a n u s c r i p t Figure 1 

COVID-19) Cases in MA As of

Development and clinical application of a rapid IgM-IgG combined antibody test for SARS-CoV-2 infection diagnosis

Analytical performance of lateral flow immunoassay for SARS-CoV-2 exposure screening on venous and capillary blood samples

Estimating SARS-CoV-2 seroprevalence and epidemiological parameters with uncertainty from serological surveys. medRxIV

Massachusetts Department of Public Health COVID-19 Dashboard -Weekly COVID-19

United States Census Bureau

Antibody responses to SARS-CoV-2 in patients with COVID-19

Quantifying antibody kinetics and RNA shedding during early-phase SARS-CoV-2 infection

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Seroprevalence of SARS-CoV-2-Specific Antibodies Among Adults

Performance Characteristics of the Abbott Architect SARS-CoV-2 IgG Assay and Seroprevalence Testing in Idaho

SARS-CoV-2 seroprevalence and neutralizing activity in donor and patient blood from the San Francisco Bay Area

Cumulative incidence and diagnosis of SARS-CoV-2 infection in New York

Known COVID-19 contact (n, %) § 47

A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t