key: cord-0776917-3ld5e7f5
authors: Lewis, Nathaniel M; Chu, Victoria T; Ye, Dongni; Conners, Erin E; Gharpure, Radhika; Laws, Rebecca L; Reses, Hannah E; Freeman, Brandi D; Fajans, Mark; Rabold, Elizabeth M; Dawson, Patrick; Buono, Sean; Yin, Sherry; Owusu, Daniel; Wadhwa, Ashutosh; Pomeroy, Mary; Yousaf, Anna; Pevzner, Eric; Njuguna, Henry; Battey, Katherine A; Tran, Cuc H; Fields, Victoria L; Salvatore, Phillip; O'Hegarty, Michelle; Vuong, Jeni; Chancey, Rebecca; Gregory, Christopher; Banks, Michelle; Rispens, Jared R; Dietrich, Elizabeth; Marcenac, Perrine; Matanock, Almea M; Duca, Lindsey; Binder, Allison; Fox, Garrett; Lester, Sandra; Mills, Lisa; Gerber, Susan I; Watson, John; Schumacher, Amy; Pawloski, Lucia; Thornburg, Natalie J; Hall, Aron J; Kiphibane, Tair; Willardson, Sarah; Christensen, Kim; Page, Lindsey; Bhattacharyya, Sanjib; Dasu, Trivikram; Christiansen, Ann; Pray, Ian W; Westergaard, Ryan P; Dunn, Angela C; Tate, Jacqueline E; Nabity, Scott A; Kirking, Hannah L
title: Household Transmission of SARS-CoV-2 in the United States
date: 2020-08-16
journal: Clin Infect Dis
DOI: 10.1093/cid/ciaa1166
sha: 8e9a945a84a5e20aa286c89f5a944b3fa32943fe
doc_id: 776917
cord_uid: 3ld5e7f5

BACKGROUND: Although many viral respiratory illnesses are transmitted within households, the evidence base for SARS-CoV-2 is nascent. We sought to characterize SARS-CoV-2 transmission within US households and estimate the household secondary infection rate (SIR) to inform strategies to reduce transmission. METHODS: We recruited laboratory-confirmed COVID-19 patients and their household contacts in Utah and Wisconsin during March 22–April 25, 2020. We interviewed patients and all household contacts to obtain demographics and medical histories. At the initial household visit, 14 days later, and when a household contact became newly symptomatic, we collected respiratory swabs from patients and household contacts for testing by SARS-CoV-2 rRT-PCR and sera for SARS-CoV-2 antibodies testing by enzyme-linked immunosorbent assay (ELISA). We estimated SIR and odds ratios (OR) to assess risk factors for secondary infection, defined by a positive rRT-PCR or ELISA test. RESULTS: Thirty-two (55%) of 58 households had evidence of secondary infection among household contacts. The SIR was 29% (n = 55/188; 95% confidence interval [CI]: 23–36%) overall, 42% among children (<18 years) of the COVID-19 patient and 33% among spouses/partners. Household contacts to COVID-19 patients with immunocompromised conditions had increased odds of infection (OR: 15.9, 95% CI: 2.4–106.9). Household contacts who themselves had diabetes mellitus had increased odds of infection (OR: 7.1, 95% CI: 1.2–42.5). CONCLUSIONS: We found substantial evidence of secondary infections among household contacts. People with COVID-19, particularly those with immunocompromising conditions or those with household contacts with diabetes, should take care to promptly self-isolate to prevent household transmission.

The Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) outbreak that began in China in December 2019 has become a global pandemic. As of July 3, 2020, the United States has reported the greatest number of cases and deaths worldwide [1] . Transmission has been reported in many settings, including health care facilities and family and community gatherings [2] [3] [4] . In China, household transmission and contact tracing investigations found that a range of 3% to 32% of household contacts acquired infection [5] [6] [7] [8] [9] [10] . However, published data on systematic household transmission investigations in US households are scarce. 

The Centers for Disease Control and Prevention collaborated (CDC) with state and local health departments in the Milwaukee, Wisconsin, and Salt Lake City, Utah, metropolitan areas to identify persons with laboratory-confirmed SARS-CoV-2 infection captured by public health surveillance during March 22-April 25, 2020. In Wisconsin, SARS-CoV-2 outpatient testing was limited to persons meeting CDC criteria for influenza testing [11] and excluded asymptomatic or mildly symptomatic A c c e p t e d M a n u s c r i p t 4 persons who were not health care workers [12] . Outpatient testing in Utah required clinical features (fever, cough, or shortness of breath) and an epidemiologic risk factor [13] . The investigation team defined persons identified by local health departments as "index patients." Households were selected by convenience sampling and considered eligible if the index patient was not hospitalized at the time, lived with ≥1 additional person, and tested positive for SARS-CoV-2 rRT-PCR from a nasopharyngeal (NP) swab collected ≤10 days prior to enrollment. All persons in the household were asked to participate; we excluded households where >1 person declined.

Prior to the first household visit, questionnaires capturing demographic characteristics, medical histories, and recent symptoms were administered to index patients and all household contacts by phone. A household-level questionnaire captured physical characteristics of the residence (e.g., square footage, number of bedrooms and bathrooms). For each household, the investigation team conducted an initial in-person visit (on day 0) and a visit 14 days later (day 14). The investigation period was defined as 14 days, the maximum duration of the SARS-CoV-2 incubation period [14, 15] .

On day 0 and day 14, we collected an NP swab, self-collected anterior nasal swab, and blood sample for all persons living in the household, including the index patient. Each person completed a daily symptom diary during days 0-14; a newly symptomatic person prompted an interim visit during which repeat swabs were obtained for all household contacts. During day 0 and day 14 visits, we interviewed households on precautionary practices to reduce transmission. Additional information regarding household selection and investigation protocols are in the Supplemental Appendix.

Swabs were tested by the Milwaukee Health Department Laboratory (MHDL) or the Utah Public Health Laboratory (UPHL) using the CDC 2019 Novel Coronavirus Real Time RT-PCR assay [16] . Blood samples were processed by MHDL or UPHL; sera were subsequently shipped to CDC and tested using a CDC-developed SARS-CoV-2 enzyme-linked immunosorbent assay (ELISA) [17] . 

For this analysis, we excluded households for which the primary patient could not be determined.

We summarized household environment characteristics, as well as demographic characteristics and medical histories of primary patients and household contacts. We estimated the SIR as the proportion of participating household contacts with secondary infection and estimated the serial interval as the number of days from symptom onset of the primary patient to symptom onset of the associated household contact with secondary infection. Statistical tests of trends were calculated using the Cochran-Armitage trend test. Two-sided p-values and 95% confidence intervals (CI) for binomial proportions were calculated with chi-square tests and Wilson score intervals, respectively.

A c c e p t e d M a n u s c r i p t 6 Characteristics of household environments, primary patients, and household contacts were assessed as potential risk factors. We estimated unadjusted odds ratios (OR) and 95% CI for potential risk factors of secondary infection among household contacts using a generalized estimation equation (GEE) approach [19] . To account for within-household correlation, we specified an exchangeable correlation structure by households. To address potential misclassification of secondary infections among household contacts, we excluded contacts with evidence of SARS-CoV-2 infections by serology only and repeated SIR calculations and risk factor analysis. We estimated risk ratios and its 95% CI for household environment characteristics as potential risk factors for whether any secondary transmission occurred in the household. We described precautionary practices reported by households on day 0 and day 14.

Data collected from questionnaires and symptom diaries were entered into a REDCap electronic database hosted at CDC [20, 21] . Analyses were conducted in SAS Enterprise Guide, version 7.1 (SAS Institute, Cary NC).

This protocol was reviewed by CDC human subjects research officials and the activity was deemed non-research as part of the COVID-19 public health response.

We enrolled sixty-two households (Utah, n = 36; Wisconsin, n = 26), and excluded four households for which we could not identify the primary patient. Among the remaining 58 households (Utah, n = 34; Wisconsin, n = 24), 58 primary patients and 188/197 (95%) household contacts were included in the analysis; 9 household contacts declined participation after their respective households were Table 2 ). The median house size was 2,200 square feet (range: 600-8,000).

Demographic and clinical characteristics of primary patients and household contacts are shown in Table 1 . The median age was 40 years (range: 16-90) for primary patients and 22 years (range: <1-76) for household contacts. Among 58 primary patients, 21 (36%) had at least one underlying medical condition; the most common symptoms were respiratory (n = 56, 97%), followed by neurologic (n = 51, 88%), constitutional (n = 50, 86%), and gastrointestinal (n = 41, 71%) (Supplementary Table 3 ). One primary patient was asymptomatic and tested based on a known nonhousehold exposure. Median intervals to the day 0 household visit were 11 days (interquartile range Table 4 ). By the initial household visit, 43/52 (83%) had acquired secondary infection; 9 (17%) household contacts acquired secondary infection during the 14-day investigation period. Eight (15%) of 52 household contacts with secondary infection had detectable SARS-CoV-2 antibodies but no positive rRT-PCR test during the investigation. Among these 8 household contacts, 3 (38%) seroconverted between days 0 and 14, 4 (50%) were seropositive on day 0, and 1 (12.5%) did not have serology testing on day 0 but was rRT-PCR-negative and seropositive on day 14; all 8

were symptomatic with symptom onset occurring ≥2 days after symptom onset of the corresponding primary patient. Table 6 ). Repeat risk factor analyses excluding household contacts who were seropositive but rRT-PCR-negative produced similar results, with the exception that household contacts of primary patients with constitutional symptoms were found to be more likely to have a secondary infection Table 8 ).

Among the 58 households, 55 (95%) reported precautionary practices at any point following symptom onset of the first laboratory-confirmed COVID-19 patient in the household; 51 (88%) households reported ill persons sleeping in a separate bedroom and 33 (57%) reported use of cloth face covers or masks by ill persons (Figure 4 ). Household-level precautionary practices assessed were not associated with preventing household transmission (Supplementary Table 8 ).

These findings from suggest that US household settings may lead to substantial SARS-CoV-2 transmission. By the day 0 household visit, 83% of the secondary infections among household contacts had already occurred, highlighting the importance of timely case identification and isolation. Transmission occurred in 55% of households and the SIR was 29% among household contacts. We demonstrate potentially increased infection risk among children and spouses of primary patients living in the same household, household contacts of primary patients with an immunocompromising condition, household contacts of male primary patients, and household contacts with diabetes mellitus.

Previously reported household SIRs from China were 3-32% [5] [6] [7] [8] [9] [10] , and a recent point-prevalence study of household transmission in New York, United States estimated a 38% SIR [22] . One study from China estimated a 28% SIR for spouses of primary patients and 4% for household contacts aged A c c e p t e d M a n u s c r i p t 10 <18 years by SARS-CoV-2 rRT-PCR testing [6] . We observed similar SIRs of 33% among spouses of primary patients and 23% among household contacts aged <18 years after limiting our results to detection by rRT-PCR alone. Our SIRs may vary from China and New York based on differences in sociocultural context and an investigation protocol (i.e., entire household testing,14-day follow-up, serologic testing) that likely captured infections missed in point-prevalence surveys, routine contact tracing, or rRT-PCR testing alone.

People with diabetes mellitus within our investigation had increased odds of acquiring SARS-CoV-2 infection, although collinearity between diabetes mellitus and obesity could have confounded the association. Among COVID-19 patients, comorbidities such as diabetes, cardiovascular, cerebrovascular, and oncologic diseases have been identified as risk factors for severe disease and increased mortality [14, [23] [24] [25] [26] [27] . Our findings illustrate that increased risk posed by diabetes for COVID-19 could include susceptibility to SARS-CoV-2 infection as well as increased morbidity and mortality, particularly among people with poorly controlled diabetes [28, 29] . People with diabetes are at increased risk for infection generally [30, 31] investigation. Reasons for this is unclear, although it may reflect behavioral differences between male primary patients and female primary patients or increased viral shedding as males have a higher likelihood of developing severe symptoms [7, 24] .

Most households demonstrated willingness to adopt precautionary practices, with most reporting ill persons sleeping in a separate bedroom and over half reporting use of cloth face covers or masks by ill persons. Although we did not find transmission differences between households where ill persons isolated and those where they did not, we were unable to capture the extent, timing, and consistency of such precautionary practices. A household transmission study from China, however, demonstrated that mask use and self-isolation decreased risk of secondary SARS-CoV-2 infection in households [8] . Further investigations are needed to identify measures that may be acceptable to and successful for US households.

These results must be considered with respect to several limitations. First, we assumed that household transmission was responsible for infections among household contacts. The household SIR could therefore be an overestimation, although concurrent stay-at-home orders should have limited community exposures. Second, misclassification of primary patients would affect risk factor analysis. Third, households were from convenience samples from 2 states, and not representative of all US households. Future studies should assess, for example, households of primary patients with severe illness and transmission dynamics in apartment buildings. Fourth, transmission had already occurred in some households by enrollment due to delays in testing and reporting at the time of the investigation. Timelier enrollment would help differentiate between transmission generations and refine our risk factor analysis among household contacts. Fifth, we may have missed infections as repeat respiratory swabs were not obtained within a 24-hour period to confirm or exclude SARS-

A c c e p t e d M a n u s c r i p t 12 period. Sixth, timing and consistency of precautionary practices were not obtained, and we were unable to evaluate their efficacy. Finally, our approach to assessing household-level risk factors for secondary transmission did not account for individual-level characteristics. Our findings should thus be considered hypothesis-generating and suitable for evaluation in future analytic studies.

Households are likely major settings of SARS-CoV-2 transmission of in the United States.

Transmission dynamics are not uniform across or within households and some people, including spouses and children of primary patients, people living with immunocompromised primary patients, and people with diabetes, may be at higher risk of secondary infection. Given public health guidance to isolate at home when sick or to quarantine at home when exposed, effective strategies to reduce 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 

COVID-19 Response Team

Utah Department of Health

Salt Lake County Health Department

City of Milwaukee Health Department

COVID-19): Cases in the

First known person-to-person transmission of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in the USA

Presymptomatic SARS-CoV-2 Infections and Transmission in a Skilled Nursing Facility

Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72 314 Cases From the Chinese Center for Disease Control and Prevention

Household transmission of SARS-CoV-2

The characteristics of household transmission of COVID-19

Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study

Household Transmission of SARS-CoV-2

Report of the WHO-China Joint Mission on Coronavirus Disease 2019 (COVID-19): World Health Organization

Household secondary attack rate of COVID-19 and associated determinants in Guangzhou, China: a retrospective cohort study. The Lancet Infectious Diseases

Guide for considering influenza testing when influenza viruses are circulating in the community

Urgent Update -Prioritization of COVID-19 Testing for Hospitalized Patients

COVID-19 Testing Criteria Process and Overview

Clinical Characteristics of Coronavirus Disease 2019 in China

Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus-Infected Pneumonia

Centers for Disease Control and Prevention. CDC 2019-Novel Coronavirus (2019-nCoV) Real-Time RT-PCR Diagnostic Panel. Available at

COVID-19): People who are immunocompromised. Available at: Validation of a SARS-CoV-2 spike protein ELISA for use in contact investigations and sero-surveillance

Models for longitudinal data: a generalized estimating equation approach

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

Characteristics and Clinical Outcomes of Adult Patients Hospitalized with COVID-19 -Georgia

Hospitalization Rates and Characteristics of Patients Hospitalized with Laboratory-Confirmed Coronavirus Disease 2019 -COVID-NET, 14 States

Host susceptibility to severe COVID-19 and establishment of a host risk score: findings of 487 cases outside Wuhan

Cerebrovascular disease is associated with an increased disease severity in patients with Coronavirus Disease 2019 (COVID-19): A pooled analysis of published literature

Cancer patients in SARS-CoV-2 infection: a nationwide analysis in China

Association of Blood Glucose Control and Outcomes in Patients with COVID-19 and Pre-existing Type 2 Diabetes

Newly diagnosed diabetes is associated with a higher risk of mortality than known diabetes in hospitalized patients with COVID-19

Risk of Infection in Type 1 and Type 2 Diabetes Compared With the General Population: A Matched Cohort Study

Diabetes and the occurrence of infection in primary care: a matched cohort study

Risk Factors for Primary Middle East Respiratory Syndrome Coronavirus Illness in Humans, Saudi Arabia

Risk Factors for MERS-CoV Seropositivity among Animal Market and Slaughterhouse Workers

Plasma glucose levels and diabetes are independent predictors for mortality and morbidity in patients with SARS. Diabetic medicine : a journal of the

Clinical features and short-term outcomes of 144 patients with SARS in the greater Toronto area

A c c e p t e d M a n u s c r i p t 14 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 16 A c c e p t e d M a n u s c r i p t