Household transmission of SARS‐CoV‐2 during the Omicron wave in Shanghai, China: A case‐ascertained study

Abstract Objectives We used a case‐ascertained study to determine the features of household transmission of SARS‐CoV‐2 Omicron variant in Shanghai, China. Methods In April 2022, we carried out a household transmission study from 309 households of 335 SARS‐CoV‐2 pediatric cases referred to a designated tertiary Children's Hospital. The detailed information can be collected from the 297 households for estimating the transmission parameters. The 236 households were qualified for estimating the secondary infection attack rates (SARI) and secondary clinical attack rates (SARC) among adult household contacts, characterizing the transmission heterogeneities in infectivity and susceptibility, and assessing the vaccine effectiveness. Results We estimated the mean incubation period and serial interval of Omicron variant to be 4.6 ± 2.1 and 3.9 ± 3.7 days, respectively, with 57.2% of the transmission events occurring at the presymptomatic phase. The overall SARI and SARC among adult household contacts were 77.11% (95% confidence interval [CI]: 73.58%–80.63%) and 67.03% (63.09%–70.98%). We found higher household susceptibility in females. Infectivity was not significantly different between children and adults and symptomatic and asymptomatic cases. Two‐dose and booster‐dose of inactivated COVID‐19 vaccination were 14.8% (5.8%–22.9%) and 18.9% (9.0%–27.7%) effective against Omicron infection and 21.5% (10.4%–31.2%) and 24.3% (12.3%–34.7%) effective against the symptomatic disease. Conclusions We found high household transmission during the Omicron wave in Shanghai due to presymptomatic and asymptomatic transmission despite implementation of strict interventions, indicating the importance of early detection and timely isolation of SARS‐CoV‐2 infections. Marginal effectiveness of inactivated vaccines against Omicron infection poses a great challenge for outbreak containment.


| INTRODUCTION
The COVID-19 pandemic caused by SARS-CoV-2 has resulted in unprecedented global health crisis and more than six million deaths worldwide since December 2019. 1 Despite the increasing natural immunity and vaccine-induced immunity are common in population, the newly emerged Omicron variant, with increased transmissibility and immune escape properties, has rapidly replaced previous strains and driven a new surge of SARS-CoV-2 infections across the world. 2,3 China maintained local containment through effective border controls and non-pharmaceutical interventions (NPIs) since 2020 and has successfully coped with several importation-linked local outbreaks of SARS-CoV-2 variants. 4 In the meantime, Chinese government spared no efforts to promote countrywide mass COVID-19 vaccination rollout among adults since April 2021 and among children aged 3-17 years since July 2021. 5,6 Nevertheless, following the first cluster of implemented to contain the outbreak, such as case isolation, contact tracing, mass testing, and city-wide lockdown. 8 As of May 31, 2022, when the lockdown was lifted, over 0.6 million confirmed cases including 588 deaths were reported in Shanghai. 9 Transmission dynamics of SARS-CoV-2 may potentially evolve over with routine SARS-CoV-2 testing regardless of symptoms to detect asymptomatic and mildly symptomatic cases. As household contacts of SARS-CoV-2-positive cases are likely to be highly exposed to the case and are known to be at high risk of infection, they are an ideal group shedding lights on SARS-CoV-2 transmission dynamics. 18 Here, we conducted a case-ascertained study to determine the features of household transmission of SARS-CoV-2 Omicron variant in Shanghai, China. In particular, we estimated the distribution of key time-to-event intervals, quantified the household transmission risk and explored the transmission heterogeneities in infectivity and susceptibility. In the meantime, we also assessed the vaccine effectiveness of inactivated COVID-19 vaccines against Omicron infection and symptomatic disease.

| Study design and participants
Between April 4 to April 27, 2022, a total of 335 SARS-CoV-2 pediatric cases from the 309 households were referred to the Children's Hospital of Fudan University, a designated hospital for management of pediatric COVID-19 cases in Shanghai ( Figure 1). All these cases were laboratoryconfirmed before hospitalization, with suspected pneumonia or comorbidities requiring special medical attention. During the outbreak, asymptomatic and milder pediatric cases were usually transferred to designated isolation facilities for medical observation. Each pediatric case was allowed to have their parents accompanying during hospital stay. Routine medical observations and PCR testing for the hospitalized children and their accompanying parents were conducted at the hospital. Other family member contacts were mandatorily required for 14-day isolation and quarantine at the community isolation facilities or centers and received PCR screening for SARS-CoV-2 every 2 days and even every day. If they developed any symptom or sign of COVID-19, an additional test was done to help timely detect infection. Cases with two consecutive RT-PCR (reverse transcription-polymerase chain reaction) negative testing results (i.e., the Cycle threshold value for SARS-CoV-2, Ct ≥ 35) were discharged from isolation. 19 We conducted a case-ascertained study to capture the information of all these 309 households of the 335 pediatric cases, including the demographics, exposures, vaccination status, infection and clinical information of both the hospitalized children and their household members (Table S1). In-depth household investigations were conducted during the hospital stay (face-to-face interview with the accompanying parents using a standard questionnaire, Table S1) and after one-week of discharge (routine telephone follow-up, to check the infection status of each household member). The households were excluded from the study if any of the cases or household members were reluctant to provide the complete information or refused to telephone follow-up. Complete information was collected from the 297 households, including 323 hospitalized children and their 951 household members ( Figure 1).
Based on the detailed information obtained from household investigations, we defined the primary case for each household as a confirmed case with a history of community exposure (i.e., exposed to SARS-CoV-2 contaminated environment or contact with a confirmed case in the community). For a household without determined source of infection, we defined the primary case as the first individual who was tested positive with RT-PCR or developed symptoms. Other household members with positive RT-PCR results were defined as secondary cases. To reduce the potential uncertainty of the results, we only focused on those households with a single primary case. After excluding households without a primary case or with co-primary cases, 239 households were eligible, including 239 primary cases and their 784 household contacts, among which 646 secondary cases were identified (Figure 1).
In this study, all the enrolled households came from the hospitalized pediatric cases of COVID-19 who was either a primary or a secondary case. To avoid overestimation of the household secondary attack rates due to the selection bias of household enrollment, the secondary infection and clinical attack rates (SARI and SARC), as well as the transmission heterogeneities and vaccine effectiveness (VE) measured based on SARI and SARC, were estimated among adult household contacts. For this reason, we further excluded three households without an adult household contact from the 239 enrolled households, and finally, a total of 236 households were qualified for analysis, including 236 primary cases and 546 adult household contacts ( Figure 1).

| Case definition and household contact
In this study, a confirmed case is defined as a person with PCRconfirmed SARS-CoV-2 infection, irrespective of clinical signs and symptoms. A symptomatic case is defined as confirmed cases who develop COVID-19-related symptoms, such as fever, cough, runny nose, sore throat, diarrhea, vomit and constitutional symptoms, and further classified as mild, moderate (non-severe pneumonia), severe and critical case based on both the national and World Health Organization (WHO) guidance, 19,20 otherwise, they will be defined as asymptomatic cases. Pneumonia was diagnosed based on either radiological evidence or typical clinical signs (fever and or cough accompanying with one of the following signs: moist rales, difficulty in breathing, fast breathing, chest indrawing). A household is defined as two or more F I G U R E 1 Flow chart describing the procedure for screening study participants people living in the same residence. A household contact is defined as any person who has resided in the same household with a confirmed case for the period from 2 days before to 14 days after the date of symptom onset or laboratory confirmation. Each hospitalized pediatric case in this study had at least one household contacts.

| Statistical analysis
We estimated the incubation period (i.e., the period of time from an exposure resulting in SARS-CoV-2 infection to symptom onset) by analyzing cases with clear exposure history. When cases reported multiple or sustained exposures, any time within their exposure windows was considered to be their possible infection time. We also estimated the serial interval (i.e., the time interval between the onset of symptoms in a primary case and his/her secondary cases), as well as the infectiousness profile (i.e., the distribution of the time interval from the onset of symptoms in a primary case to the infection in his/her secondary cases). For a secondary case contacts with multiple infections, we randomly selected one as his/her primary case and simulated 100 times to account for potential uncertainties (see Sun. et al for more details). 21 A sensitivity analysis for the situation that all secondary cases are from the same primary cases was also conducted.
We fitted three parametric distributions (Weibull, gamma, and lognormal) to time-to-event data and selected the best fit based on the minimum Akaike information criterion. The distributions of serial interval and the infectiousness profile were fitted with a shift parameter allowing negative values.
We further excluded the households without a primary case or with co-primary cases from the analysis to avoid potential bias, as it is possible that a secondary case may be misclassified as a co-primary case ( Figure 1). The secondary infection attack rate (SAR I ) was defined as the number of PCR-confirmed cases detected regardless of symptom among all household contacts of the primary case. 18 The secondary clinical attack rate (SAR C ) was defined as the number of symptomatic cases detected among all household contacts of the primary case. 18 In this study, there was a potential bias in the estimates of SAR I and SAR C among children household contacts due to the study design that the enrolled households were selected from the families of the confirmed hospitalized pediatric cases. Therefore, we estimated the SAR I and SAR C among adult household contacts to assess the heterogeneities in infectivity and susceptibility. Specifically, the heterogeneities in susceptibility were estimated by the characteristics (e.g., sex and vaccination status) of adult household contacts.
The heterogeneities in infectivity were measured by the characteristics (e.g., age, sex, household size, symptom profile, and vaccination status) of primary cases (including children and adults).
Comparison between groups was performed using chi-square test. A difference with P < 0.05 at two-side was considered to be statistically significant. We estimated the vaccine effectiveness against Omicron infection (VE I ) and against clinical symptoms (VE C ) based on the estimates of SAR I and SAR C among adult household contacts with different vaccination status. Specifically, the estimates of VE I were

| RESULTS
A total of 297 households, including 323 hospitalized pediatric cases and their 951 household members, were initially recruited to the study ( Figure 1). The epidemiological and clinical characteristics were described in Table S2 and Figure S1. All household cases in this study were non-severe or asymptomatic, except a 7-year-old child, who was critically ill. The mean time interval between first enrollment and last follow-up was 20.9 days for each household.
We  (Table S3). We estimated the distribution of serial interval and the infectiousness profile based on the 234 transmission pairs (accounting for the uncertainty caused by co-primary cases). The serial interval followed a best fitted Weibull distribution with an estimated mean of 3.9 (median: 4.0, COVID-19-related symptoms IQR 1.4-6.4) days and a standard deviation of 3.6 days ( Figure 2B and Table S4), while the infectiousness profile followed a gamma distribution with 57.2% of the transmission events occurred at the presymptomatic phase ( Figure 2C and Table S5). Sensitivity analysis (for the situation that all secondary cases of a household are from the same primary case, not accounting for the uncertainty caused by co-primary cases) showed similar results, with an estimated serial interval of 3.8 ± 3.8 days and 53.3% of the household transmission occurred at the presymptomatic phase ( Figure S2). In this study, detailed information on exposures and symptoms of the study participants was collected through in-depth household investigations, allowing us to provide robust estimation of several key time-to-event distributions. We observed a mean incubation period of 4.6 ± 2.1 days for Omicron variant, slightly longer than prior estimates for Omicron (3.0-3.6 days) [23][24][25][26][27] while shorter than that of the ancestral strain (6.3 days). 28  Households without a primary case or with co-primary cases were excluded from this analysis. We assessed the susceptibility among adult household contacts to avoid potential bias due to the study design (detailed in Section 2).
T A B L E 2 Infectivity of primary cases, measured by secondary infection attack rate (SAR I ) and secondary clinical attack rate (SAR C ), based on the analysis of 236 primary cases and their 546 adult household contacts a Households without a primary case or with co-primary cases were excluded from this analysis. We assessed the infectivity of primary cases among their adult household contacts to avoid potential bias due to the study design (detailed in Section 2). b The SAR I and SAR C among adult household contacts of a partially vaccinated primary case were not estimated due to the extremely small sample size (i.e., 12 household contacts corresponding to seven primary cases).
babysitters. On the other hand, the enrolled pediatric cases were almost younger children, thus, the grandparents were not very old.
We estimated the overall SAR I among adult household contacts to be 77.11%, around 2.5-6 times higher than previous estimates ( During the Omicron wave, substantial increase in pediatric cases of COVID-19 was reported in the United States. 42 However, the role of children in Omicron transmission has yet to be fully understood. We observed similar high infectivity in pediatric cases (aged 0-17 years) and in adults (aged 18+ years), indicating that children played an equal role in Omicron transmission in household as adults.
Our finding also demonstrated the similar high-level transmission rate from symptomatic and asymptomatic primary cases, which implies that symptom-based surveillance is insufficient to prevent and control of COVID-19 epidemic, posing great challenge for prevention and control of Omicron transmission. We found females were more susceptible to Omicron infection in household than males, in line with the finding reported in an early study from Wuhan. 11

ACKNOWLEDGMENTS
We thank Professor Hongjie Yu for his critical comments on the manuscript. We also thank all participants for their collaboration with the investigation.

CONFLICT OF INTEREST
All the authors declared no conflicts of interest related to this work.

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/irv.13097.

DATA AVAILABILITY STATEMENT
The data and codes that support the findings of this study are available from the corresponding author upon reasonable request.