key: cord-322385-sc2vxxnn authors: Ebinger, J.; Botwin, G. J.; Albert, C. M.; Alotaibi, M.; Arditi, M.; Berg, A. H.; Binek, A.; Botting, P. G.; Fert-Bober, J.; Figueiredo, J. C.; Grein, J. D.; Hasan, W.; Henglin, M.; Hussain, S. K.; Jain, M.; Joung, S.; Karin, M.; Kim, E. H.; Li, D.; Liu, Y.; Luong, E.; McGovern, D. P. B.; Merchant, A.; Merin, N. M.; Miles, P. B.; Minissian, M.; Nguyen, T.-T.; Raedschelders, K.; Rashid, M. A.; Riera, C. E.; Riggs, R. V.; Sharma, S.; Sternbach, S.; Sun, N.; Tourtellotte, W. G.; Van Eyk, J. E.; Sobhani, K.; Braun, J. G.; Cheng, S. title: SARS-CoV-2 Seroprevalence Across a Diverse Cohort of Healthcare Workers date: 2020-08-04 journal: nan DOI: 10.1101/2020.07.31.20163055 sha: doc_id: 322385 cord_uid: sc2vxxnn Importance: Antibody testing is important for understanding patterns of exposure and potential immunity to SARS-CoV-2. Prior data on seroprevalence have been subject to variations in selection of individuals and nature as well as timing of testing in relation to exposures. Objective: We sought to determine the extent of SARS-CoV-2 seroprevalance and the factors associated with seroprevelance across a diverse cohort of healthcare workers. Design: Observational cohort study of healthcare workers, including SARS-CoV-2 serology testing and participant questionaires. Participants: A diverse and unselected population of adults (n=6,062) employed in a multi-site healthcare delivery system located in Los Angeles County, including individuals with direct patient contact and others with non-patient-oriented work functions. Exposure: Exposure and infection with the SARS-CoV-2 virus, as determined by seropositivity. Main Outcomes: Using Bayesian and multi-variate analyses, we estimated seroprevalence and factors associated with seropositivity and antibody titers, including pre-existing demographic and clinical characteristics; potential Covid-19 illness related exposures; and, symptoms consistent with Covid-19 infection. Results: We observed a seroprevalence rate of 4.1%, with anosmia as the most prominently associated self-reported symptom in addition to fever, dry cough, anorexia, and myalgias. After adjusting for potential confounders, pre-existing medical conditions were not associated with antibody positivity. However, seroprevalence was associated with younger age, Hispanic ethnicity, and African-American race, as well as presence of either a personal or household member having a prior diagnosis of Covid-19. Importantly, African American race and Hispanic ethnicity were associated with antibody positivity even after adjusting for personal Covid-19 diagnosis status, suggesting the contribution of unmeasured structural or societally factors. Notably, number of people, or children, in the home was not associated with antibody positivity. Conclusion and Relevance: The demographic factors associated with SARS-CoV-2 seroprevalence among our healthcare workers underscore the importance of exposure sources beyond the workplace. The size and diversity of our study population, combined with robust survey and modeling techniques, provide a vibrant picture of the demographic factors, exposures, and symptoms that can identify individuals with susceptibility as well as potential to mount an immune response to Covid-19. diagnosis status, suggesting the contribution of unmeasured structural or societally factors. Notably, number of people, or children, in the home was not associated with antibody positivity. The demographic factors associated with SARS-CoV-2 seroprevalence among our healthcare workers underscore the importance of exposure sources beyond the workplace. The size and diversity of our study population, combined with robust survey and modeling techniques, provide a vibrant picture of the demographic factors, exposures, and symptoms that can identify individuals with susceptibility as well as potential to mount an immune response to Covid-19. Amidst the ongoing global pandemic caused by SARS-CoV-2, the viral agent causing Covid-19, substantial attention 1 turned to antibody testing as an approach to understanding patterns of exposure and immunity across populations. The use and interpretion of antibody testing to assess exposure and immunity remains frought with inconsistencies and unclear clinical correlations, in part due to a dearth of high quality studies among diverse participants. 2, 3 Recent publications have pointed to the challenges and importance of understanding how different antibody tests for SARS-CoV-2 perform, and factors that may render one method superior to another. 4, 5 Nonetheless, there remains general agreement that antibody testing offers valuable information regarding the probable extent of SARS-CoV-2 exposure, the factors associated with exposure, and the potential nature and determinants of seropositive status. 6 To the end, we conducted a study of SARS-CoV-2 antibody screening of a large, diverse, and unselected population of adults employed in a multi-site healthcare delivery system located in Los Angeles County, including individuals with direct patient contact and others with nonpatient-oriented work functions. Recognizing the range of factors that might influence antibody status in a given individual, we focused our study on not only estimating seroprevalence but also on identifying factors associated with seropositivity and relative antibody levels within the following three categories: (1) pre-existing demographic and clinical characteristics; (2) potential Covid-19 illness related exposures; and, (3) Covid-19 illness related response variables (i.e. different types of self-reported symptoms). The sampling strategy for our study has been described previously. 7 In brief, beginning on May 11, 2020, we enrolled a total of N=6,318 active employees working at multiple sites comprising the Cedars-Sinai Health System, located in the diverse metropolis of Los Angeles County, California. The Cedars-Sinai organization includes two hospitals (Cedars-Sinai Medical Center and Marina Del Rey Hospital) in addition to multiple clinics in the Cedars-Sinai Medical Delivery Network. All active employees (total N~15,000) were invited to participate in the study by providing a peripheral venous blood sample for serology testing and completing an electronic survey of questions regarding past medical history, social history, and work environment in addition to Covid-19 related symptoms and exposures. 8, 9 For the current study, we included all participants who completed both SARS-CoV-2 antibody testing and electronic survey forms (N=6,062). The study protocol was approved by the Cedars-Sinai institutional review board and all participants provided written informed consent. All participant biospecimens underwent serology testing by the Cedars-Sinai Department of Pathology and Laboratory Medicine using the Abbott Diagnostics SARS-CoV-2 IgG chemiluminescent microparticle immunoassay assay (Abbott Diagnostics, Abbott Park, IL) performed on an Abbott Diagnostics Architect ci16200 analyzer. The assay reports a signal-tocutoff ratio (S/CO) corresponding to the relative light units produced by the test sample compared to the relative light units produced by an assay calibrator sample. The manufacturer recommended S/CO ratio of 1.4 was used to assign binary seropositivity status. This cutoff was validated for high specificity (i.e., >99%) ~14 days post symptom onset. 10 The Abbott assay detects antibodies directed against the nucleocapsid (N) antigen of the SARS-CoV-2 virus, which assists with packaging the viral genome after replication, and achieves specificity for IgG by incorporating an anti-human IgG signal antibody. Estimates of Seroprevalence. We conducted a literature review to identify published data (until June 25, 2020) on the sensitivity and specificity of the Abbott Architect SARS-CoV-2 IgG assay, applied in specific populations using the manufacturer's recommended thresholds. We identified a total of 15 studies assessing sensitivity in 2,114 tests and 18 studies reporting specificity in 7,748 tests (Supplemental Tables 1-2); we combined this information with data from an additional independent cohort of 60 case and 178 control specimens used to asses sensitivity and specificity, respectively, within the Cedars-Sinai Department of Pathology and Laboratory Medicine. We noted that studies investigating specificity generally assessed samples collected prior to the SARS-CoV-2 pandemic whereas studies reporting sensitivity included specimens from RT-PCR confirmed individuals (see details provided in Supplemental Tables 1-2) . We restricted our analyses to a referent cohort of tests conducted on samples from individuals who were assayed ≥ 7 days following symptoms onset to most closely match our cohort sample characteristics and the situational context for study enrollment. We integrated source population-level demographic data, representative of the entire Cedars-Sinai employee base, with data from our enrolled study sample using an Iterative Proportional Fitting procedure (IPF) to estimate the number of eligible employees within each demographic category (with provided population totals considered the target, using constraints derived from our sample). 11 We then fit a Bayesian multilevel hierarchical logistic regression model using RStan, 12,13 including reported age, gender, race/ethnicity and site as coefficients, to model exposure probability (see Supplemental Methods for full details). We estimated the seroprevalence within each post-stratified demographic category based on the averaged and weighted value of the expected number of employees within that category. Factors Associated with Seroprevalence. Prior to multivariable-adjusted analyses, age and IgG index were transformed by dividing by 10 for interpretability of coefficients in all models. In adjusted analyses, we compared differences between serology status (i.e. antibody positive versus negative) in each variable of interest, grouped into one of three categories: (1) preexisting demographic and clinical characteristics (e.g. age, gender, ethnicity, race, and selfreported medical comorbidities); (2) Covid-19 related exposures (e.g. self-reported medical diagnosis of Covid-19 illness, household member with Covid-19 illness, number of people living in the home including children, type of home dwelling, etc); and, (3) Covid-19 related response variables (e.g. self-reported fever, chills, dry cough, anosmia, nausea, myalgias, etc.). In multivariable-adjusted analyses, we used logistic and linear models to examine the extent to which the three categories of variables (predictors) may be associated with antibody positive status (primary outcome) in the total sample or IgG antibody level in the subset of persons with positive antibody status (secondary outcome). Initial models were deliberately sparse, adjusting for a limited number of key covariates (e.g. age, gender) and those variables with associations meeting a significance threshold of P<0.10 were advanced for inclusion in a final multivariable model with only other variables identified from the sparse regression included. A final separate multivariable model was constructed for each of the 3 categories of variables. The demographic, clinical, exposure, and symptom response characteristics of the study sample are shown in Table 1 , by antibody test result status; the study sample included individuals whose residence spanned diverse regions across Los Angeles County (Supplemental Figure 1) . The overall seroprevalence was 4.1% (95% CI 3.1%, 5.7%), with higher estimates seen in younger compared to older individuals and in Hispanics compared to non-Hispanics (Figure 1 and Supplemental Table 3 ). In multivariable-adjusted analyses of pre-existing characteristics (Figure 2 and Supplemental In multivariable-adjusted analyses of Covid-19 related exposures (Figure 3 and Supplemental In multivariable-adjusted analyses of Covid-19 response variables (Figure 4 and Supplemental Table 6 ), the strongest self-reported symptom associated with greater odds of seropositive status was anosmia (11.53 [7.51, 17. Table 7 ). In a single combined model including significant pre-existing, exposure, and symptom factors, the main findings were unchanged from those in the primary analyses (Supplemental Figure 2 ). In a large diverse healthcare employee cohort of over 6,000 adults in Los Angeles, we observed a seroprevalence rate of 4.1%, which when accounting for published test characteristics, may range from 3.1% to 5.7%. Seroprevalence varied across demographic, clinical, exposure and symptom based characteristics. Specifically, factors significantly associated with presence of IgG antibodies included younger age, Hispanic ethnicity, and African-American race, as were exposure related factors including the presence of either a personal or household member having a prior medical diagnosis of Covid-19. Among self-reported symptoms, anosmia was most strongly associated with the presence of antibodies, with positive associations also noted for fever, dry cough, anorexia, and myalgias. The size and diversity of this study population, combined with robust survey and modeling techniques, provide a more vibrant picture of the population at highest risk for Covid-19 infection, risks of various potential exposures and symptoms that should alter patients to potential illness. Most prior seroprevalence studieshave focused on cohorts that included healthcare workers predominantly involved in direct or indirect patient care, persons living within a circumscribed region with high viral exposure rates, or larger geographic areas from which motivated individuals could voluntarily enroll into community screening programs. 14, 15 Given that completely unbiased population-scale sampling for seroprevalence studies remains a logistical challenge, we used a sampling approach that involved open enrollment and convenient access to testing facilities made available to all employees working across multiple sites of a large healthcare system; this approach was intended to broadly capture individuals with both patientrelated exposures and community-related exposures, while also representative of a relatively wide geographic area in and around Los Angeles County. Although limited to persons who are generally healthy and able to be employed, our study cohort included individuals representing a diversity of demographic characteristics including ethnicity and race -leading to findings that reflect the disparities that have been persistently observed and reported for Covid-19 infection rates in our local communities. Consistent with findings from studies in healthcare workers, seroprevalence patterns in our cohort indicate exposure from not only the work environment but also from the home environment and likely unmeasured community-based factors. 16 It has been well reported that minority populations, particularly African Americans and Hispanics, have been disproportionately effected by the Covid-19 panedmic. [17] [18] [19] Our study is consistent with these prior findings, but demonstrates that such differences exist even when all participants work not just in the same field, but for the same organization. Such a finding may indicate that community and non-work related environmental factors are likely playing a significant role in the spread of Covid-19 among certain minority populations. Even after controlling for a medical diagnosis of Covid-19, African American race and Hispanic ethnicity remained risk factors for antibody positivity. The persistence of thse racial and ethnic disparities may represent structural barriers to care or societally mediated risk. Geographic clustering by race and ethnicity in housing, shopping and social gatherings may be one such factor, while socioeconomic status and ability to self-isolate outside of work likely also contribute. [20] [21] [22] No self-reported pre-existing medical conditions were significantly associated with antibody positivity, indicating that infection itself is agnostic to baseline health. In fact, asthma was negatively associated with the presence of antibodies, or at least antibody levels above the current threshold we use for positivity. While reactive airway disease is unlikely a protective factor against Covid-19, participants with such conditions may be more likely to deligently follow social distancing guidelines and practice better adherence to hand hygiene and use of personal protective equipment. Hypertension was the only medical condition associated with higher SARS-CoV-2 antibody levels. It remains unclear as to what physiologic mechanism may contribute to this finding, however, unmeasured confounding variables, such as medications or renal disease may function as mediating factors. Further studies will be needed to both verify and elucidate this finding. Also concordant with prior studies, we found that anosmia was the single strongest symptom associated with SARS-CoV-2 IgG antibody presence. [23] [24] [25] Interestingly, neither dyspnea nor diarrhea, two commonly cited symptoms, demonstrated a significant association in multivariable analysis. 26, 27 This is likely related to the non-specific nature of these symptoms, which are common to multiple viral and non-viral etiologies. Importantly, dyspnea was associated with a higher antibody level among those with anti-SARS-CoV-2 antibodies, suggesting that dyspnea related to Covid-19 may drive a more robust humoral immune response, potentially related to more severe infection. These findings are concordant with the known phenomenon of proportionate adaptive immune response to higher doses of antigenic stress. 28 The extent to which the generation of measurably higher antibody levels could confer immunity to a larger degree or for a longer duration of time remains unknown. Interestingly, prior studies have demonstrated lower antibody levels among exposed, asymptomatic individuals, a phenomena which may be attributable to a highly efficient cell mediated immune response. 29 It has be suggested that higher T-cell levels, whether virus specific or otherwise, may play a role in this finding, however, further research is required. 30, 31 Further expanding from prior studies, we investigated and observed several factors that appeared notably unassociated with seroprevalence. In particular, we found that recent travel, type of home, and number of people living in the home were not associated with an antibodybased measure of SARS-CoV-2 exposure. The presence of antibodies was also not related to youth or children in the home, or to having domestic pets such as cats or dogs. Although far from definitive, these results suggest that these factors do not play an important role in mediating potentially meaningful viral exposure in the communities represented by our study cohort. Several limitations of this study merit consideration. Of the employees actively employed at our multi-site institution, only a proportion of all eligible participants enrolled; nonetheless, the sample size of the cohort was large, diverse, and representative of the source sample. 7 Our seroprevalence estimates were based on using a validated assay of only IgG antibodies; assays of IgM antibodies may offer complementary information in future studies. Data collected on medical history, exposures, and symptoms were all self-reported, similar to approaches used in prior studies. We were unable to completely verify prior Covid-19 illness using viral test results in part given lack of universally available testing for all individuals, particularly those with minimal to no symptoms. In conclusion, in a highly diverse population of healthcare workers, demographic factors associated with Covid-19 antibody positivity indicate potential factors outside of the workplace associated with SARS-CoV-2 exposure, although these do not appear related to the number of people or to the presence of children in the home. Further, while for dyspnea may be a marker of more severe disease among those with Covid-19, it's presence alone does not indicate infection. Foundation. 175 Included inpatients and outpatients PCR positive from >= 8 PSO Kohmer et al. 38 35 45 From 45 pts with positive PCR Stroemer et al. 39 33 34 34 sera obtained from 26 patients between four and 60 days (median 19 days) after a positive real-time RT-PCR. Nicol et al. Models are adjusted for significant predictors from the primary multivariable models examining associations of existing characteristics, exposures and symptoms with antibody positivity and IgG titer index. Supplemental Figure 1 . Model 1 is adjusted for age, sex, race, ethnicity Serology for SARS-CoV-2: Apprehensions, opportunities, and the path forward Policy for Diagnostic Tests for Coronavirus Disease-2019 during the Public Health Emergency SARS-CoV-2 infection serology: a useful tool to overcome lockdown? Developing antibody tests for SARS-CoV-2 Will antibody tests for the coronavirus really change everything? Review of Current Advances in Serologic Testing for Covid-19 An Opportune and Relevant Design for Studying the Health Trajectories of Healthcare Workers. medRxiv Research electronic data capture (REDCap)-A metadata-driven methodology and workflow process for providing translational research informatics support The REDCap consortium: Building an international community of software platform partners Performance Characteristics of the Abbott Architect SARS-CoV-2 IgG Assay and Seroprevalence in mipfp: An R Package for Multidimensional Array Fitting and Simulating Multivariate Bernoulli Distributions A Probabilistic Programming Language The prevalence of severe acute respiratory coronavirus virus 2 (SARS-CoV-2) IgG antibodies in intensive care unit (ICU) healthcare personnel (HCP) and its implications-a singlecenter, prospective, pilot study Prevalence of IgG antibodies to SARS-CoV-2 among emergency department employees Hospital-Wide SARS-CoV-2 Antibody Screening in 3056 Staff in a Tertiary Center Racial Health Disparities and Covid-19 -Caution and Context Covid-19 by Race and Ethnicity: A National Cohort Study of 6 Million United States Veterans. medRxiv The Disproportionate Impact of Covid-19 on Racial and Ethnic Minorities in the United States Examining Social Determinants of Health, Stigma, and Covid-19 Disparities The Structural and Social Determinants of the Racial/Ethnic Disparities in the U.S. Covid-19 Pandemic: What's Our Role? Disparities in the Population at Risk of Severe Illness From Covid-19 by Race/Ethnicity and Income Olfactory and gustatory dysfunctions as a clinical presentation of mild-to-moderate forms of the coronavirus disease (Covid-19): a multicenter European study The Prevalence of Olfactory and Gustatory Dysfunction in Covid-19 Patients: A Systematic Review and Meta-analysis Self-reported anosmia and dysgeusia as key symptoms of coronavirus disease Clinicopathological characteristics of 8697 patients with Covid-19 in China: a meta-analysis Clinical Insights into the Gastrointestinal Manifestations of Covid-19 Efficacy of High-Dose versus Standard-Dose Influenza Vaccine in Older Adults Clinical and immunological assessment of asymptomatic SARS-CoV-2 infections Targets of T Cell Responses to SARS-CoV-2 Coronavirus in Humans with Covid-19 Disease and Unexposed Individuals Phenotype and kinetics of SARS-CoV-2-specific T cells in Covid-19 patients with acute respiratory distress syndrome ARCHITECT SARS-CoV-2 IgG Instructions for Use SARS-CoV-2 seroprevalence and neutralizing activity in donor and patient blood from the San Francisco Bay Area. medRxiv Highthroughput immunoassays for SARS-CoV-2, considerable differences in performance when comparing three methods. medRxiv SARS-CoV-2 Antibody responses do not predict Covid-19 disease severity Clinical evaluation of serological IgG antibody response on the Abbott Architect for established SARS-CoV-2 infection Performance Characteristics of Four High-Throughput Immunoassays for Detection of IgG Antibodies against SARS-CoV-2 Brief clinical evaluation of six high-throughput SARS-CoV-2 IgG antibody assays Diagnostic accuracy of six commercial SARS-CoV-2 IgG/total antibody assays and identification of SARS-CoV-2 neutralizing antibodies in convalescent sera Assessment of SARS-CoV-2 serological tests for the diagnosis of Covid-19 through the evaluation of three immunoassays: Two automated immunoassays (Euroimmun and Abbott) and one rapid lateral flow immunoassay (NG Biotech) Evaluation of Covid-19 IgG/IgM Rapid Test from Orient Gene Biotech Side by side comparison of three fully automated SARS-CoV-2 antibody assays with a focus on specificity Sensitivity of commercial Anti-SARS-CoV-2 serological assays in a high-prevalence setting. medRxiv Clinical Performance of Two SARS-CoV-2 Serologic Assays Diabetes is a risk factor for the progression and prognosis of Covid-19. Diabetes/metabolism research and reviews Performance of six SARS-CoV-2 immunoassays in comparison with microneutralisation Validation and Performance Comparison of Three SARS-CoV-2 Antibody Assays Patients with Common Cold Coronaviruses Tested Negative for IgG Antibody to SARS-CoV-2 We are grateful to all the front-line healthcare workers in our healthcare system who continue to be dedicated to delivering the highest quality care for all patients.