key: cord-319770-ddq2q1pg
authors: Wen, Ying; Wei, Lan; Li, Yuan; Tang, Xiujuan; Feng, Shuo; Leung, Kathy; Wu, Xiaoliang; Pan, Xiong-Fei; Chen, Cong; Xia, Junjie; Zou, Xuan; Feng, Tiejian; Mei, Shujiang
title: Epidemiological and clinical characteristics of COVID-19 in Shenzhen, the largest migrant city of China
date: 2020-03-23
journal: nan
DOI: 10.1101/2020.03.22.20035246
sha: 
doc_id: 319770
cord_uid: ddq2q1pg

We conducted a retrospective study among 417 confirmed COVID-19 cases from Jan 1 to Feb 28, 2020 in Shenzhen, the largest migrant city of China, to identify the epidemiological and clinical features in settings of high population mobility. We estimated the median incubation time to be 5.0 days. 342 (82.0%) cases were imported, 161 (38.6%) cases were identified by surveillance, and 247 (59.2%) cases were reported from cluster events. The main symptoms on admission were fever and dry cough. Most patients (91.4%) had mild or moderate illnesses. Age of 50 years or older, breathing problems, diarrhea, and longer time between the first medical visit and admission were associated with higher level of clinical severity. Surveillance-identified cases were much less likely to progress to severe illness. Although the COVID-19 epidemic has been contained in Shenzhen, close monitoring and risk assessments are imperative for prevention and control of COVID-19 in future.

The Coronavirus Disease 2019 (COVID-19) caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) is emerging as a major infectious disease epidemic globally. Initially detected in a cluster of patients with unexplained pneumonia in Wuhan, Hubei Province of China in early December 2019, SARS-CoV-2 rapidly spread not only within China but also around the globe within just three months. As of Mar 10, 2020, 113,702 confirmed cases and 4,012 deaths have been reported in 109 countries (1) . Although the natural reservoir of SARS-CoV-2 is still unknown, early confirmed cases are strongly associated with exposures to wild animals in the Huanan Seafood Wholesale Market (2) , and sustained human-to-human transmission is observed particularly among close contacts (3, 4) .

Due to the surging number and rapid spread of COVID-19, World Health Organization (WHO) has increased the risk assessment of COVID-19 to 'very high' at the global level on Feb 28, 2020.

As one of the most developed and commercialized cities in China, Shenzhen is the largest migrant city where over 80% of its population (20 million) are migrants. The risk of case importation in Shenzhen was therefore high, especially because the COVID-19 epidemic occurred around the Chinese Lunar New Year holiday season during which intra-city mobility of the migrant population was extremely high. Since the first case of COVID-19 in Shenzhen was confirmed on Jan 20, 2020, local authorities activated the highest level of emergency response to the disease. Although there have been several studies on the transmission, epidemiology and clinical symptoms of COVID-19 in Wuhan (3, 5, 6) , data from other epidemic areas are still lacking (7) . In particular, information and knowledge of COVID-19 from a migrant city with high population mobility like Shenzhen can inform effective prevention and control strategies in other similar settings. As such, we investigated the epidemiological and clinical characteristics of all 417 cases that were confirmed in Shenzhen is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. . https://doi.org/10.1101/2020.03.22.20035246 doi: medRxiv preprint as of Feb 28, 2020.

We conducted a retrospective study of the epidemiological and clinical characteristics of COVID-19 cases in Shenzhen from Jan 1 to Feb 28, 2020. COVID-19 case was defined in accordance with the WHO interim guidelines and the National Guidelines in Diagnosis and Treatment Scheme for COVID-19 (Sixth edition) (8, 9) . Sputum, blood, broncho-alveolar lavage fluid, nasopharyngeal swab, or oropharyngeal swab were collected from each patient and tested using real-time polymerase chain reaction (RT-PCR) for SARS-CoV-2. A cycle threshold value less than 37 was defined as PCR test positive. Virus detection was first done by hospitals or districted centers for disease control and prevention (CDC), and positive results were further confirmed by Guangdong Provincial CDC (before Jan 30, 2020) or Shenzhen CDC (after Jan 30, 2020). For each confirmed case, detailed epidemiological investigations were conducted by local epidemiologists and public health workers. All confirmed cases were immediately reported to the National Infectious Disease Information System for COVID-19, which was amended as a Class B notifiable infectious disease on Jan 20, 2020.

We compiled the epidemiological and clinical data from both local epidemiological investigation reports and the National Infectious Disease Information System. Patients' information including sociodemographic characteristics, exposure history, close contacts, time-lines of illness onset, medical visit, hospitalization, and PCR confirmation, symptoms, and clinical outcomes, was extracted to construct a dataset with no personal identity. All confirmed COVID-19 cases in Shenzhen were included in the current analyses with no prespecified exclusion criteria. Data collection in the epidemiological investigation was part of the continuing public health investigation of an emerging outbreak and therefore the . CC-BY-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

Patients were defined as having Wuhan exposure if they were Wuhan residents or visited Wuhan within the past 14 days before symptom onset, and were defined as having Shenzhen exposure if they had not left Shenzhen within the past 14 days before symptom onset, while all others exposed elsewhere were defined as having exposure elsewhere in mainland China other than Wuhan or Shenzhen. Based on the National Guidelines in Diagnosis and Treatment Scheme for COVID-19 (9), the seriousness of clinical presentations of COVID-19 cases was categorized as mild, moderate, severe, and critical. Mild cases were those with virological confirmation but without an evidence of having pneumonia. Moderate cases were mild cases at the same time with a diagnosis of pneumonia. Severe was defined when one of the following criteria was met: dyspnea (respiratory frequency≥30/minute); blood oxygen saturation ≤93%; PaO2/FiO2 ratio <300; and/or lung infiltrates >50% within 24-48 hours.

Critical cases were defined when patients had respiratory failure; and/or septic shock; and/or multiple organ dysfunction/failure. In addition, we categorized cases into self-identified and surveillance-identified ones. The former referred to COVID-19 cases who were identified when they sought medical care at the hospitals, while the latter referred to cases who were identified through active surveillance efforts including screening of close contacts of the confirmed patients and recent travelers from Hubei; fever monitoring at airport, train station, docks, and highway checkpoints; and registration and report of fever by community workers.

We defined a cluster as occurrence of two or more confirmed cases in a socially-close setting (such as a family, a school, or a company) within the past 14 days, which may be caused by human-to-human transmission through close contacts within such setting or infection via a common external exposure. Other cases not from a cluster were defined as scattered cases. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

Demographic and clinical characteristics as categorical variables were presented as numbers and percentages, while continuous variables were presented as means and standard deviations or 95% confidence intervals (CI), or medians and interquartile ranges (IQR) if appropriate. Inter-group differences in the characteristics were tested by using Pearson's χ 2 test or Fisher's exact test for categorical variables, and by using Student's t-test or analysis of variance for continuous variables showing a normal distribution, and Kruskal-Wallis and Wilcoxon tests for continuous variables with non-parametric distribution. The incubation period was estimated by using a previously described parametric accelerated failure time model (10). Patients with detailed information on the time of exposure, the date of illness onset, or the first time of presentation were included for this analysis. We fitted lognormal, gamma, and Weilbull distributions using Markov Chain Monto Carlo in a Bayesian framework (11). We estimated the serial interval by using the time difference of illness onset between the infector and infectee. Initial reproductive number was estimated by using the best fit model based on date of illness onset of the early (Jan 10-23) local exposed cases without relation to the imported cases and the estimated serial interval of COVID-19. Logistic regression models were applied to identify factors associated with the clinical severity of COVID-19. All statistical tests and analyses of the incubation period, serial interval, and initial reproductive number were performed in R software (R foundation for Statistical Computing).

A total of 417 cases had been confirmed as of Feb 28, 2020 in Shenzhen. In terms of potential source of exposure, 224 (53.7%) were exposed in Wuhan, 75 (18.0%) were exposed in Shenzhen, and 118 (28.3%) were exposed elsewhere (Table 1 ). In addition, 161 (38.6%) cases were identified by surveillance, and 247 (59.2%) cases occurred in 92 clusters (common is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. . https://doi.org/10.1101/2020.03.22.20035246 doi: medRxiv preprint exposure or secondary transmission). None of the patients had been to the Huanan Seafood Wholesale Market within the past 14 days before illness onset, which was initially thought to be the index location of zoonotic infections of SARS-CoV-2 that started the COVID-19 epidemic (12). days assuming the incubation time followed a lognormal distribution. Estimates from models with other distributions (gamma and Weilbull) were 5.4 to 5.5 days (See supplemental materials). Using information on the date of illness onset from 28 pairs of infector and infectee, we estimated the mean serial interval to be 5.5 (95% CI: 4.1-7.0) days. We estimated the initial reproductive number to be 2.5 (95% CI: 1.4-4.3) using information from the cases that exposed in Shenzhen.

Demographic characteristics, clinical symptoms and outcomes were presented in Table 1 .

The mean age was 45.4 years old, and 220 (52.8%) cases were female. Patients with Wuhan exposure were older than those with exposure in Shenzhen or elsewhere (48.2 vs 42.9, 41.6, . CC-BY-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. . https://doi.org/10.1101/2020.03.22.20035246 doi: medRxiv preprint P<0.05). Patients identified by surveillance were younger than those self-identified cases (40.5 vs 48.4, P<0.001). In total, there were 29 (7.0%) pediatric patients aged below 15 years, of whom 26 were identified by surveillance and 28 were cases from clusters.

The most common symptoms of COVID-19 were fever (67.4%), dry cough (34.3%), and myalgia (28.5%). Patients who were exposed in Wuhan (78.1%), identified by surveillance (83.6%), or scattered (78.2%) were more likely to have fever (P<0.001). Patients with Shenzhen exposure were less likely to have dry cough (21.3%) than patients with Wuhan exposure (35.7%) and elsewhere exposure (39.8%); they were also less likely to have myalgia (26.7%) than those with Wuhan exposure (33.5%). Patients who were identified by surveillance were less likely to have symptoms than the self-identified, such as fever (41.6% vs 83.6%), dry cough (23.0% vs 41.4%), myalgia (10.6% vs 39.8%), sore throat (8.1% vs 19.1%), and headache (8.1% vs 16.4%). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

In this study, we reported the epidemiological and clinical characteristics of confirmed COVID-19 cases in Shenzhen, a large migrant city with the highest GDP per capita in mainland China. Our results can be used to improve the prediction of transmission risk, design and implementation of intervention measures and strategies, and assessments of intervention effectiveness in similar settings.

The epidemic curve of COVID-19 in Shenzhen had practical implications. Overall, the rapid increase of cases in Shenzhen was interrupted by the massive control measures implemented since Jan 23 and followed by a sustained downward trend. The majority of the cases were imported and the local transmission was limited. Furthermore, most cases were self-identified at the beginning, while cases were more likely to be identified by surveillance is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. Fever and cough were the most common symptoms of COVID-19 on hospital admission in Shenzhen, which was in accordance with findings from a study that included 1099 patients from 30 provinces of China (14) . This county-wide study also reported that the proportion of patients with fever increased (88.7%) after hospital admission, although the proportion (43.8%) on admission was lower than that in our study (67.4%) (14). The overall proportion of patients with cough was lower than that of other study (59.4%) (5); meanwhile, more patients had dry cough (34.3%) than cough with phlegm (15.1%) in our study. The main symptoms were significantly different regarding different exposure sources; patients with Shenzhen exposure were less likely to have fever and cough compared to those with Wuhan exposure. In addition, the patients in our study were less likely to have symptoms compared to those patients reported in Wuhan (15). Interestingly, surveillance-identified patients showed fewer symptoms at the time of identification than those self-identified patients in our study.

This might be because active surveillance identified cases at an early stage of COVID-19.

Our study reported preliminary findings on the clinical severity of COVID-19 in is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Other studies also suggested that these factors were associated with poorer outcomes in patients in Wuhan and elsewhere (5, 15) . Furthermore, the longer time from the first medical visit to hospital admission was associated with higher clinical severity on admission, while surveillance identification was associated with lower clinical severity. Thus the low proportion of severe or critical cases in our study might be explained by the large number of cases identified by surveillance and the short time interval from the first medical visit to hospital admission. Of note, 33 (83.3%) of the severe or critical cases were selfidentified, which might be due to delayed hospital admission. Through active surveillance efforts such as screening of high-risk populations, a large proportion of COVID-19 patients were identified at the early stage of their illness, thus decreasing the possibility of progression to a severe illness. The hospital fatality rate of COVID-19 in Shenzhen (0.7%) was much lower than 14% reported from Wuhan (16), and also lower than 1.4% reported in a countywide study (15). Since data on clinical outcomes were censored in our study, the hospital fatality rate may be underestimated.

Although our study had major strengths such as population-wide case identification in a major migrant city, a large sample size, and complete profiling of epidemics along the timeline of population interventions, there were several limitations that should be acknowledged. First, since some patients were still under treatment so far, we could not yet report the complete data of the disease progression on severity or assess predictors of clinical outcomes. Second, important information, such as timelines of possible exposure, illness onset, and medical visits, was self-reported in the epidemiological investigations, which might be subject to recall bias. Third, we did not have enough information on comorbid conditions, is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint 

(70.0) *Elsewhere exposed refers to cases have been exposed in cities in mainland China other than Wuhan or Shenzhen.

. CC-BY-ND 4.0 International license It is made available under a perpetuity.

is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint .47) *AOR, adjusted OR; the model is constructed using the stepwise logistic regression (entry probability 0.05; removal probability 0.10) with variables including age, fever, cough with phlegm, sore throat, headache, breath problem, diarrhea, days from illness onset to the first medical visit, days from the first medical visit to admission, identification mode, and cluster. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. . https://doi.org/10.1101/2020.03.22.20035246 doi: medRxiv preprint Figure 1 The epidemic curve of COVID-19 cases by the date of illness onset and PCR confirmation in Shenzhen. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

The copyright holder for this this version posted March 23, 2020. . https://doi.org/10.1101/2020.03.22.20035246 doi: medRxiv preprint

Novel Coronavirus (2019-nCoV) situation reports Genava2020

A new coronavirus associated with human respiratory disease in China

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

The 2019-nCoV Outbreak Joint Field Epidemiology Investigation Team QL. An Outbreak of NCIP (2019-nCoV) Infection in China-Wuhan

Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Clinical findings in a group of patients infected with the 2019 novel coronavirus (SARS-Cov-2) outside of Wuhan, China: retrospective case series

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