key: cord-0749219-xsqgrd5l
authors: Cheng, Hao-Yuan; Jian, Shu-Wan; Liu, Ding-Ping; Ng, Ta-Chou; Huang, Wan-Ting; Lin, Hsien-Ho
title: High transmissibility of COVID-19 near symptom onset
date: 2020-03-19
journal: nan
DOI: 10.1101/2020.03.18.20034561
sha: 915059b415f823b526d4fe1bdf952094f91bde5e
doc_id: 749219
cord_uid: xsqgrd5l

Background The dynamics of coronavirus disease 2019 (COVID-19) transmissibility after symptom onset remains unknown. Methods We conducted a prospective case-ascertained study on laboratory-confirmed COVID-19 cases and their contacts. Secondary clinical attack rate (considering symptomatic cases only) was analyzed for different exposure windows after symptom onset of index cases and for different exposure settings. Results Thirty-two confirmed patients were enrolled and 12 paired data (index-secondary cases) were identified among the 1,043 contacts. The secondary clinical attack rate was 0.9% (95% CI 0.5-1.7%). The attack rate was higher among those whose exposure to index cases started within five days of symptom onset (2.4%, 95% CI 1.1-4.5%) than those who were exposed later (zero case from 605 close contacts, 95% CI 0-0.61%). The attack rate was also higher among household contacts (13.6%, 95% CI 4.7-29.5%) and non-household family contacts (8.5%, 95% CI 2.4-20.3%) than that in healthcare or other settings. The higher secondary clinical attack rate for contacts near symptom onset remained when the analysis was restricted to household and family contacts. There was a trend of increasing attack rate with the age of contacts (p for trend < 0.001). Conclusions High transmissibility of COVID-19 near symptom onset suggests that finding and isolating symptomatic patients alone may not suffice to contain the epidemic, and more generalized social distancing measures are required. Rapid reduction of transmissibility over time implies that prolonged hospitalization of mild cases might not be necessary in large epidemics.

The coronavirus disease 2019 (COVID-19) outbreak originated from Wuhan has spread to more than 100 countries only two months after the virus, Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), was identified in January 2020. 1, 2 Following the Wuhan lockdown and other extreme social distancing measures conducted by the Chinese government, several countries with massive outbreaks also implemented similar measures, including shutting down the entire cities or communities, banning international or domestic travel, conducting border control with symptom screening, and implementing isolation and quarantine to slow down the epidemic.

However, for a novel pathogen like SARS-CoV-2, its unknown epidemiologic characteristics and transmission dynamics complicated the development and evaluation of effective control policies. 3 Researches on COVID-19 have sprouted with growing epidemics in different countries and provided some valuable insights. The short transmission cycle (serial interval) of COVID-19 and results from viral shedding studies suggested the possibility of transmission near or even before symptom onset, while prolonged viral shedding raised concerns about prolonged infectiousness and the need for extended hospital stay. [4] [5] [6] A few preliminary contact-tracing studies showed that the highrisk exposure setting of COVID-19 transmission was in the household. [7] [8] [9] Nevertheless, these fragmented knowledges were still inadequate to answer some practical questions like, when and how long we should isolate a COVID-19 patient or quarantine close contacts. To connect these insights and reveal the full picture of COVID-19 transmission, evidence from the field is urgently needed to provide information about the transmission risk at different time points after symptom onset and at different exposure settings.

Taiwan's first COVID-19 case was confirmed on January 21. 10 With proactive containment efforts and comprehensive contact tracing, the number of COVID-19 cases in Taiwan was maintained low compared to other countries overwhelmed by massive outbreaks. 11 Using the contact tracing data in Taiwan, we aimed to delineate the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 19, 2020 

In response to the outbreak in Wuhan, Taiwan Centers for Disease Control (Taiwan CDC) made COVID-19 as a notifiable disease on January 15, 2020. A prospective caseascertained study enrolling all confirmed cases and their close contacts was conducted between January 15 and February 26.

A confirmed case should meet the criteria of notification for COVID-19 in Taiwan and be tested positive by real-time reverse transcriptase-polymerase chain reaction (RT-PCR) test. 12 Detailed information including demographic and clinical data was reported to the National Notifiable Disease Surveillance System. 13 The investigation team determined the clinical severity of the confirmed patients following the World Health Organization (WHO) interim guidance. 14

When the patient was laboratory-confirmed to have SARS-CoV-2 infection, a thorough epidemiological investigation including contact tracing was implemented by the outbreak investigation team of Taiwan CDC and local health authorities. The definition of a close contact was a person who did not wear appropriate personal protection equipment (PPE) while having face-to-face contact with a confirmed case for more than 15 minutes after symptom onset. A contact was listed as a household contact if he/she lived in the same household with the index case, while those listed as family contacts were family members not living in the same household. For health care settings, close contact was defined by contacting an index case within two meters without appropriate PPE. Medical staff, hospital workers, and other patients in the same setting were included.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 19, 2020. . https://doi.org/10.1101/2020.03. 18.20034561 doi: medRxiv preprint All close contacts were quarantined at home for 14 days after their last exposure to the index case. During the quarantine period, any relevant symptoms (fever, cough, or other respiratory symptoms) of close contacts would trigger RT-PCR testing for COVID-19.

For high-risk populations, including household and hospital contacts, RT-PCR was performed regardless of symptoms. In Taiwan CDC, we used an electronic tracing system (Infectious Disease Contact Tracing Platform and Management System) to follow and record the daily health status of those quarantined contacts. The information collected included age, gender, the index case, date of exposure, and the exposure setting.

Paired data of index case and close contacts were extracted from the contact tracing database and outbreak investigation reports. For a family cluster, the index case was determined based on the temporality of symptom onset and review of the epidemiological link. A secondary case was excluded from the paired data if the date of onset was earlier than the date of exposure. Similarly, a close contact would be excluded if the date of last exposure was earlier than the date of symptom onset of the index case. For health care contacts, the date of exposure would be the date of admission if the exact date of exposure was not recorded.

Incubation period and serial interval were estimated using the contact tracing data and publicly available datasets (Supplement Appendix). We used the Bayesian hierarchical model to increase the stability in small-sample estimation. The exposure window period was defined as the time period from the first contact after symptom onset of the index case to the end of contact. For household contacts who lived with the index case, the day of first contact was set to be the day of symptom onset of the index case. Secondary clinical attack rate was calculated by dividing the number of symptom confirmed cases by the number of close contacts. We analyzed the dynamic change of secondary clinical attack rate after symptom onset in several ways. First, we categorized contacts based on the time of the initial exposure to the index case after index case's symptom onset (days 0-3, 4-5, 6-7, 8-9, or >9). Second, we considered the whole exposure window period and fitted logistic regression models to estimate the piecewise period-specific odds ratio . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . under different definitions of exposure (binary exposure or cumulative person-time exposure in each categorical period). Third, we estimated the density function of transmission at different timing of exposure since symptom onset, considering the whole exposure window period and assuming a gamma distribution of the density function (See Supplementary appendix for details of the statistical analysis).

The data management and analysis were done using R software (R Foundation for Statistical Computing) and RStan (Stan Development Team).

By February 26, there were 32 laboratory-confirmed COVID-19 patients, including five household/family clusters and four asymptomatic patients. Sixteen (50%) patients were imported, and the remaining 16 cases were locally acquired. Of the 16 locally-acquired cases, three patients without any travel history were detected because of pneumonia of unknown etiology. The remaining 13 cases were secondary cases found through contact tracing. One of the 13 cases was excluded from subsequent transmission pair analysis because the documented day of exposure occurred after symptom onset of the secondary case. The median delay from symptom onset to confirmed diagnosis was 5 days (range 0-12) among imported cases and 18.5 days (range 2-28) among locally-acquired cases ( Figure 1A ). All secondary cases had their first day of exposure within five days of the index case's symptom onset ( Figure 1B ). The presumed transmission trees were depicted in Figure 1C . We estimated that the mean incubation period was 4.9 days (95% credible interval [CrI] 2.7-8.4), and the mean serial interval was 7.0 days (95% CrI 3.7-13.2).

A total of 1,043 close contacts were identified. Among them, 3.4% were household contacts, 4.5% were non-household family contacts, and 28.9% were health care contacts ( Table 1 ). The risk for COVID-19 infection (considering 12 transmitted cases) in the contacts was 1.2% (95% confidence interval [CI] 0.6-2.0%), and the secondary clinical attack rate was 0.9% (95% CI 0.4-1.6%) among all contacts. The clinical attack rate was higher among those whose initial exposure to the index case was within five days of symptom onset (2.4%, 95% CI 1.1-4.5%) than those who were exposed later (zero . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . transmission out of 605 contacts, 95% CI 0-0.61%) ( Table 2 and Figure 2A ). The attack rate from early exposure remained high when we restricted the analysis to household and non-household family contacts (Table 3 and Figure 2B ). Further analyses that accounted for the whole exposure window period also revealed higher transmissibility near symptom onset ( Figure 2C-2D) .

The secondary clinical attack rate was 13.6% (95% CI 4.7-29.5%) among household contacts and 8.5% (95% CI 2.4-20.3%) in non-household family contacts ( Table 2) . No nosocomial infection was observed in this study. There was a trend of increasing secondary clinical attack rate with the age of contacts, ranging from 0% (95% CI 0-5.1%) in those aged less than 20 years to 3.6% (95% CI 0.8-10.3%) in those aged 60 years and above (p for linear trend: <0.001). Close contacts of confirmed cases presented with mild illness were not at a lower risk compared to the contacts of more severe cases.

The secondary attack rate among contacts of locally-acquired cases was higher than that among contacts of imported cases (Table 2 ).

This study is one of the very few initial reports of the secondary clinical attack rate is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . https://doi.org/10.1101/2020.03. 18.20034561 doi: medRxiv preprint Our analysis suggested that the shorter serial interval of COVID-19 was due to the combination of early-stage transmission and a short period of infectiousness.

The observed pattern of secondary clinical attack rate over time was also consistent with the quantitative data of the SARS-CoV-2 viral shedding in upper respiratory specimens, which reported a high viral load around the time of symptom onset, followed by a gradual decrease in viral shedding to a low level after 10 days. 5 The viral load was similar among asymptomatic, minimally symptomatic, and symptomatic patients. Another virological study in COVID-19 patients also found no viable isolates of the virus after the first week of symptoms. Our data agreed with the virological data on high transmissibility in the first week and much decreased risk afterwards. 16 To summarize the evidence, the decreasing risk for secondary infection over time in our study, the observed short serial interval, and the trend of decreasing viral shedding and viability after symptom onset strongly suggested high transmissibility of the disease near or even before the day of symptom onset. Since the onset of overt clinical symptoms such as fever, dyspnea, and signs of pneumonia usually occurred 5-7 days after initial is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . https://doi.org/10.1101/2020.03. 18.20034561 doi: medRxiv preprint when the transmissibility has started to decrease. In this case, hospitalization would not be helpful for isolation and reducing transmission, and should be spared for severe patients. When the number of confirmed cases increases rapidly, home care for patients with mild illness should be considered. 20 In Taiwan, the most prolonged duration of hospital isolation for the 32 confirmed cases was more than one month. If every patient with mild illness is to be isolated in the hospital or other isolation institutes for such a prolonged period, the healthcare system will soon be overwhelmed during a large epidemic. In this case, an unusually high case fatality rate such as that observed in Wuhan may occur. 21, 22 Similarly, better understanding of the potential duration of transmission could help direct containment strategies. For example, the efforts of contacting tracing could focus on the contacts near or even before symptom onset of the index cases when the number of index cases or contacts is too large to afford. To fight a potential pandemic like COVID-19, improved efficiency for resource allocation will be critical since a massive outbreak will rapidly consume public health and medical resources.

Several patients in our study had pneumonia with unknown etiology and had multiple contacts in the healthcare setting before being diagnosed, but none of these healthcare contacts resulted in nosocomial transmission. Besides the basic PPE used by medical staffs, this result might be due to their late admissions and lower transmissibility by the time of hospitalization as well. This pattern is compatible with the observation in China and Hong Kong. In China, the number of nosocomial infections might be lower than that reported because some healthcare workers acquired infections in their households rather than in the healthcare setting; 9 In Hong Kong, most hospitalization was also delayed to at least 5 days after disease onset. 23 In closed settings like hospital or cruise ship, 24 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . https://doi.org/10.1101/2020.03. 18.20034561 doi: medRxiv preprint Our study has limitations. First, we did not completely examine contacts before the symptom onset of the index cases. Therefore, we might have underestimated the importance of early transmission. In other words, the actual importance of early transmission could be higher than our estimates. Our findings agree with the recommendation from WHO on using four days before symptom onset as the starting date for contact tracing. 28 

Information collection were done by the pronouncement of the Central Epidemic Command Center (CECC), and, in accordance with Article 17 of the Communicable Disease Control Act. As part of the public health response functions of the CECC for surveillance purposes, institutional review board approval of this study was waived. The data was deidentified prior to analysis.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . https://doi.org/10.1101/2020.03. 18.20034561 doi: medRxiv preprint bureaus (Taipei city, New Taipei city, Taoyuan city, Changhua county, and Kaohsiung county) for their dedicated outbreak investigation and meticulous date collection. Our works could not be done with their efforts. We also thank Chia-Lin Lee for the development of electronic contact tracing system (Epidemic Intelligence Center, Taiwan CDC); Ching-Hung Wang (Tony Q) for the consultation of system development; Angela Song-En Huang for writing assistance.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . ARDS, acute respiratory distress syndrome.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted March 19, 2020. . . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 19, 2020 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted March 19, 2020. . https://doi.org/10.1101/2020.03. 18.20034561 doi: medRxiv preprint 

A Novel Coronavirus from Patients with Pneumonia in China

World Health Organization. Coronavirus disease 2019 (COVID-19) Situation Report -48 [Internet]. World Health Organization

Defining the Epidemiology of Covid-19 -Studies Needed

Serial interval of novel coronavirus (COVID-19) infections

SARS-CoV-2 Viral Load in Upper Respiratory Specimens of Infected Patients

Viral load of SARS-CoV-2 in clinical samples

Community Transmission of Severe Acute Respiratory Syndrome Coronavirus 2

2019-nCoV CDC Response Team. Initial Public Health Response and Interim Clinical Guidance for the 2019 Novel Coronavirus Outbreak -United States

Report of the WHO-China Joint Mission on Coronavirus Disease

First case of Coronavirus Disease 2019 (COVID-19) pneumonia in Taiwan

Response to COVID-19 in Taiwan

Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR

International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity

Real-Time Surveillance of Infectious Diseases: Taiwan's Experience

Clinical management of severe acute respiratory infection when novel coronavirus (2019-nCoV) infection is suspected

World Health Organization

Transmission Dynamics and Control of Severe Acute Respiratory Syndrome

Clinical presentation and virological assessment of hospitalized cases of coronavirus disease 2019 in a travel-associated transmission cluster

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan, China

Feasibility of controlling COVID-19 outbreaks by isolation of cases and contacts. The Lancet Global Health

Home care for patients with suspected novel coronavirus (COVID-19) infection presenting with mild symptoms, and management of their contacts [Internet]. World Health Organization

Backcalculating the Incidence of Infection with COVID-19 on the Diamond Princess

Surface Environmental, and Personal Protective Equipment Contamination by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) From a Symptomatic Patient

Environmental Contamination and Viral Shedding in MERS Patients During MERS-CoV Outbreak in South Korea

Extensive Viable Middle East Respiratory Syndrome (MERS) Coronavirus Contamination in Air and Surrounding Environment in MERS Isolation Wards

The First Few X (FFX) Cases and contact investigation protocol for 2019-novel coronavirus (2019-nCoV) infection [Internet]. World Health Organization

We thank the Taiwan COVID-19 outbreak investigation team (as listed in Supplement Appendix), and staffs of regional control centers of Taiwan CDC and public health