key: cord-0829923-9bz7enek
authors: Li, Xia; Wang, Qin; Ding, Pei; Cha, Yu’e; Mao, Yixin; Ding, Cheng; Gu, Wen; Wang, Youbin; Ying, Bo; Zhao, Xiaoning; Pan, Lijun; Li, Yunpu; Chang, Junrui; Meng, Congshen; Zhou, Jun; Tang, Zhigang; Sun, Ruofeng; Deng, Fuchang; Wang, Chong; Li, Li; Wang, Jiao; MacIntyre, C. Raina; Wu, Zunyou; Feng, Zijian; Tang, Song; Xu, Dongqun
title: Risk factors and on-site simulation of environmental transmission of SARS-CoV-2 in the largest wholesale market of Beijing, China
date: 2021-03-02
journal: Sci Total Environ
DOI: 10.1016/j.scitotenv.2021.146040
sha: 6709e5009a9b17ce402aa61bd5cf0303f9fe8442
doc_id: 829923
cord_uid: 9bz7enek

From June 11, 2020, a surge in new cases of coronavirus disease 2019 (COVID-19) in the largest wholesale market of Beijing, the Xinfadi Market, leading to a second wave of COVID-19 in Beijing, China. Understanding the transmission modes of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and the personal behaviors and environmental factors contributing to viral transmission is of utmost important to curb COVID-19 rise. However, currently these are largely unknown in food markets. To this end, we completed field investigations and on-site simulations in areas with relatively high infection rates of COVID-19 at Xinfadi Market. We found that if goods were tainted or personnel in market was infected, normal transaction behaviors between sellers and customers, daily physiological activities, and marketing activities could lead to viral contamination and spread to the surroundings via fomite, droplet or aerosol routes. Environmental factors such as low temperature and high humidity, poor ventilation, and insufficient hygiene facilities and disinfection practices may contribute to viral transmission in Xinfadi Market. In addition, precautionary control strategies were also proposed to effectively reduce the clustering cases of COVID-19 in large-scale wholesale markets.

On June 11, 2020, after 56 days without any new diagnoses of COVID-19 cases, a second wave suddenly sparked in Beijing, China that gripped the whole world's attention 1, 2 . The center of this outbreak, Xinfadi Wholesale Market, is the largest food market in Asia, and was thus suspended immediately, and strict control measures (e.g., closed the market, locked down certain residential compounds, closed all schools, and canceled hundreds of flights)

were quickly implemented to contain the spread. Between June 15 and July 10, individuals deemed as high-risk (e.g., close contacts of confirmed cases and residents living nearby to the market) were tested by nucleic acid test using qRT-PCR, with over 10 million personnel and 5,342 environmental samples tested overall. A total of 401 confirmed cases and asymptomatic cases had been reported, of which 368 cases [169 (46.0%) employees and 103 (28.0%) visitors of Xinfadi Market as well as 96 (26.2%) contacted with these infected employees or visitors] were in Beijing, and 33 cases were in other provinces linked with this outbreak, suggesting a single outbreak source in Beijing. On August 6, the last hospitalized patient was discharged from hospital in Beijing, and no death had occurred.

This coronavirus resurgence was linked to a food market, which is reminiscent of the first outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that was identified at a Huanan Seafood Wholesale Market in Wuhan, Hubei Province. Cluster epidemics have also been reported within seafood industries 3 and meat processing facilities 4 worldwide (e.g., in the United States, Germany, Australia and France). Previous epidemiological investigations and viral genome sequencing of Xinfadi strains speculated that 6 

The site of this investigation and simulation experiment was the Basement One of Xinfadi Wholesale Market, Fengtai District, Beijing where most of cases (122/169, 72.2%) among Xinfadi employees was found. Basement One is a sunken underground space with an area of approximately 10,000 square meters and 5 meters height surrounded by circular corridor. The entrance is located in the north and south of Basement One. There are several hundreds of business booths selling aquatic products, soy products, beef, lamb, chicken, etc.

Each booth has a unified counter, and drainage channel is located around the counter. Two toilets (around 55m 2 ) for male and female are on the south. According to the findings of epidemiological investigation during outbreak 5 , seller cases were mainly distributed on the west side of Basement One with the incidence rates of 51.5% and 28.0% in aquatic and soy products areas, respectively 2, 5 . Before the experiments, on-site investigation was conducted to understand the layout floor design (e.g., floor plans, air-conditioning system, ventilation routes, drainage system, sanitation facilities, and toilet) of Basement One. The aquatic and soy products areas plus high-risk toilet were finally chosen for the following simulation experiments ( Figure 1A) .

Fluorescent powders (Beijing Cypress Chemical Co., LTD., China) with diameter 48 μm are colorless under natural light but showed red fluorescence under blacklight (365 nm). This J o u r n a l P r e -p r o o f 7 powder did not exist in nature and were therefore used as a tracer in on-site experiments.

Fluorescent polystyrene latex microspheres with a diameter 2 μm were provided by Section of Ecological Environment and Energy Resources at Beijing Institute of Metrology. It showed silver-gray under scanning electron microscope (SEM) and green under blacklight or fluorescence microscope. As a proxy for SARS-CoV-2, SARS-CoV-2 spike pseudovirus, which has no autonomous replication ability, was purchased (Sino Biological Inc., Beijing, China).

Before the on-site experiments, the dispersion differences between SARS-CoV-2 spike pseudovirus and fluorescent microsphere were compared in water, aerosol and object surface at laboratory. For water experiment, a same amount of pseudovirus (10 9 copy/mL) and fluorescent microsphere (10 9 balls/mL) was respectively prepared in tap water and then serial diluted with tap water. For aerosol and object surface experiments, pseudovirus and fluorescent microsphere were respectively placed in a micro-aerosol generator to simulate a sneezing and then contaminated the plates. The sprayed aerosols were immediately collected, while the plates were exposed to aerosols for 30 mins and then smeared for analysis.

RT-PCR (Rotor-Gene Q, QIAGEN, MD, USA) was used to measure viral loads in all collected samples. Fluorescent probe method was used to detect the copy number of the WPRE gene of pseudovirus, and the obtained CT value was converted to obtain the copy number in each sample. Qiagen's QIAamp® Viral RNA Mini Kit was used to extract the J o u r n a l P r e -p r o o f 8 genome, which serves as a template to reverse-transcribe the RNA using the cDNA first-strand synthesis kit (Sino Biological Inc., Beijing, China). 1 μL cDNA was taken for RT-PCR detection with a detection limit of 7.97E+03 virus copies/mL. Fluorescence microscope (Nikon Corporation, Tokyo, Japan) were used to measure the number of fluorescent microspheres in all collected samples. Our results indicated that the dispersions of pseudovirus and fluorescent microspheres in water, aerosols and object surfaces were consistent and insignificant ( Figure S1 ). Therefore, fluorescent microspheres could replace viruses with a consistent transmission dynamic and better traceability in the following simulation experiments.

The market was strictly controlled after the outbreak, and we were allowed to complete the experiments in a day and could not recruit real sellers. Therefore, both buyers and sellers were the members of our CDC team, who have well learnt the behaviors of sellers and buyers by observing, consulting and and mimicking from real sellers and buyers in other food markets. After the completion of terminal disinfection of the market, on-site simulation experiments were carried out on July 8, 2020. During the experiments, the indoor air-temperature was 19-20°C and the relative humidity was 55.5%, the atmospheric pressure was 100.4 kPa, and the central air-conditioning system was turned on. All the CDC staffs wore full personal protective equipment (PPE), namely an N95 respirator, goggles, gloves, shoe covers, cap, and protective apparel. The stands of ultraviolet lamps (40W) were settled J o u r n a l P r e -p r o o f 9 around the experiment areas of Xinfadi. During the experiment, blacklight were turned on all the time, and the track of fluorescent powders with red fluorescence was captured by a HD camera. Fluorescent microsphere solution was freshly prepared to reach 10 8 balls/mL with 1 hr vortex to avoid aggregation, and 1-2 mL of fluorescent solution was sprayed to tainted goods (e.g., live fish and mock tofu) to simulate the contamination from a sneezing or coughing of an infected cases (e.g., seller) during the simulation experiments.

For the aquatic products area, a live fish was sprayed with the fluorescence solution to simulate the "tainted fish". The common transactions and activities between a seller and a customer were simulated surrounding A1-1 counter (Figure 2A) , including choosing live fish, killing, weighing, removing entrails, scraping fish scales, payment, and transferring. After this, the actions of cleaning the countertop with clean water and cleaning rag was simulated. There was one pair of simulated seller and customer for this area, and all the processes were repeated twice and lasted for 30 mins at the same area within the seller's booth.

For the soy products area, water-absorbing paperboard instead of tofu was designed as the "tainted tofu" contaminated by fluorescence solution. The transactions between a seller and a customer were simulated surrounding S1 counter (Figure 2B) , including cutting, weighing, calculation, and trading. The actions of cleaning the countertop with clean water and cleaning rag were also mimicked. There was one pair of simulated seller and customer for this area, and all the processes were repeated twice and lasted for 20 mins at the same area within the seller's booth. After the business hours, the ground was washed with pressure J o u r n a l P r e -p r o o f washer gun daily by the market cleaner. Hence, this process was mimicked in both aquatic products and soy products areas and lasted for 10 mins.

For the toilet, a semi-solid solution of 0.8% agar broth mixed with fluorescence solution was used to simulate "tainted feces" from an infected person. The activities at toilet, including waiting, flushing, walking around, and leaving, were simulated. All the processes were repeated twice and lasted 30 mins.

For all the three simulated areas, long handle cotton swabs (Copan Medical Equipment Co., LTD, Shanghai, China) were used to collect surface samples (e.g., chopping board, knife handle, blade, fishing net, calculator, countertop, door handle, ground, and PPEs) before (as blank control) and after the experiments (e.g., sales transactions, washing the countertop, and washing the ground), The smear sampling area was 5cm×5cm and repeated three times. After collection, they were immediately stored in sealed plastic bags and kept away from light, and then transferred to the laboratory for further analysis. Before, during or after the simulation experiments, the sterile plates with 9 cm diameter (Jindian Biochemical Equipment Co. LTD, Qingdao, China) were placed at different distances and heights within 15 m of operation sites according to the droplet diffusion distances of larger droplets and aerosols, and each simulation scenario was collected 20 mins～2 hr for the particles (droplets) precipitation in air 11 . After sampling, the plates were sealed and stored away from light and transferred to laboratory for analysis. 

Non-parametric Wilcoxon test was conducted to determine whether there is a statistically significant difference among different groups using R software version 3.3.1 (R Development Core Team, 2006). A p value less than 0.05 was considered as statistically significance. Bar plot was visualized using ggplot2 package in R.

J o u r n a l P r e -p r o o f 13 The high COVID-19 infection rates of 51.5% and 28% were found for employees from aquatic products and soy products areas in the Basement One of Xinfadi Market, respectively ( Figure 1A ) 5 . This is further supported by the high positive rates (7.9% and 15.2%, respectively) for SARS-CoV-2 of surface samples from objects and goods in these two areas ( Figure 1A) . Fan-coil air conditioners with fresh air were used in Basement One during the outbreak. Since the counters of business booth are open, the end air outlets and the air return inlets can circulate air between the nearby counters (Figures S2). There were four air-conditioning units to exchange the air of Basement One. However, the inward air came from the circular corridor instead of outdoor fresh air, and hence the fresh air exchange rate was low. It was not only poorly ventilated, but it also caused unsterilized and unfiltered air to circulate in nearby counters. The drainage channels were opened to the air around all the counters, which was only about 15 cm wide and 5 cm deep. Objects (e.g., fish scales and debris) often blocked these shallow drainage channels, resulting in a poor drainage and humid environment. There were one entrance and one window in the toilets with 8 cubicles and two hand-washing taps, and no any hand hygiene tools (e.g., soap and sanitizer) and disinfection (e.g., UV light) were available.

Before the simulation experiments, smear and sedimentation samples were all negative from the aquatic products area, soy products area, and toilet. Detailed results are shown in Supplementary Appendix Tables S1-S3.

For the aquatic products area, in brief, after the trading processes, all of touchable items (e.g., fishnets, knife, chopping board, platform scale button, calculator, etc.) were positive, and fluorescent microspheres could be detected in smear samples with a positive rate of 65.0% (13/20), but most of sedimentation samples were negative (Figure 2A and Table S1 ).

After the washing the countertop, the positive rate of the sedimentation samples was 62.5% After flushing the toilet, in the cubicle of simulated area, smear samples from the ground next to squat toilet, wastebasket in cubicle T2, and some surfaces of water tank were positive, whereas the surrounding wall and the wall above the water tank were negative ( Figure 2C and Table S3 

To the best of our knowledge, this is the first on-site simulation experiment attempting to explain the potential spread patterns of SARS-CoV-2 in a real clustering epidemic center using fluorescent powders and microspheres as tracers. Through on-site investigations and simulating common behaviors and activities (e.g., sales transactions, fish slaughter, countertop washing, ground washing, defecating, toilet flushing, and walking around) that may cause viral spread in the market, samples of environment, object surfaces and PPEs were collected to explore the potential routes and environmental factors of SARS-CoV-2 transmission in Xinfadi market. In general, the results indicated that if goods were tainted or personnel in the market were infected by SARS-CoV-2, normal sales transaction behaviors between sellers and customers (e.g., scaling, fish slaughter, weighing, delivering, payment, etc.), daily physiological activities (e.g., using toilet, talking, walking, etc.), and marketing activities (e.g., countertop-washing, ground-washing, etc.) could easily contribute contamination and spread to the surroundings through fomite, droplet or aerosol routes.

Most confirmed cases and positive environmental samples in Xinfadi Market appeared along the aisles, which connect aquatic products area and soy products area ( Figure 1A) . μm aerodynamic diameter 9 . Another study found SARS-CoV-2 particles with sizes >4 μm and 1-4 μm in two AIIRs rooms 14 . In the present study, the particle sizes of fluorescent microspheres used were about 2 μm. Results indicated that these particles could mimic virus-containing aerosols, and are small enough to remain suspended in air for a long period of time and be inhaled 9 .

It is worth noting that positive sedimentation samples were detected away from experiment site in aquatic area and as long as 2 hrs after the completion of experiment in toilet. Besides, fluorescent microspheres were detected in aerosol samples collected after counter-washing, ground washing, and toilet-flushing. Previously several studies have also reported that SARS-CoV-2 were found on toilets and in the air in settings such as hospitals and apartment used by COVID-19 patients 7, 9, 14, 15 . SARS-CoV-2 has been frequently detected in feces and urine [16] [17] [18] [19] [20] , which can be aerosolized through toilet flushing 21 . During toilet flushing, massive upward transport of virus aerosol particles was observed, with 40-60% of particles rising above the toilet seat 22 . Therefore, toilets are a daily necessity but may promote aerosol transmission if used improperly during COVID-19 pandemic. Furthermore, fluorescent microspheres were detected on the surface of PPEs such as N95 masks, googles, and protective apparels that worn by our staff responsible for the conduction of simulation experiments, sample collections and on-site coordination, further confirming the plausibility of droplets (<1 m to the contamination source) and aerosols (>1 m to the contamination source) transmissions. This is in accordance with a previous study showing SARS-CoV-2 virus on protective apparel or in medical staff areas 23 , which was found to be resuspended as a source of aerosols by the movements of medical staffs. Therefore, activities including walking and handling goods (e.g., scraping fish scales and washing the good surface) may resuspend particles on surfaces and floors and then trigger surface aerosolization in the market, which has been demonstrated by the 7.7% of the aerosol samples were positive in the aquatic J o u r n a l P r e -p r o o f products areas. Altogether, evidence indicated aerosol transmission could not be ruled out during the outbreak of Xinfadi.

The dissemination of viruses across the market may be influenced by some environmental factors. First, low temperature and high humidity may enable the virus to viable for a long-term. Though limited laboratory data so far, results indicated that SARS-CoV-2 was relatively stable at low temperature 24 . The temperature of Basement One was low, especially for the counters selling frozen and chilled products. Counters selling fresh aquatic products has accumulated water around the surrounding area during the trading period, and the aisles were wet because of drainage block and walking activities. SARS-CoV-2 has been frequently detected in wastewater in many countries and significantly correlated with the increase in reported COVID-19 prevalence [25] [26] [27] , indicating that the virus survives in humid environments longer than expected 25, 28 . In contrast, viral activity is greatly reduced after drying 29, 30 . In the case of high humidity, the drying time of droplets is about 5 times longer than that of low humidity, which may greatly increase the survival time of virus in the environment 31 . In addition, the survival rate of SARS-CoV-2 on the surface in high humidity environment is slightly higher than that in medium humidity 32, 33 .

Second, the ventilation of Basement One is poor. It is a sunken underground space inside a building, surrounded by corridors. There are four air-conditioning units to exchange the air.

However, the inward air came from the circular corridor instead of outdoor fresh air, hence J o u r n a l P r e -p r o o f the fresh air exchange rate was low. Since the counters of business booth were open, the end of fan-coil air conditioners' air outlets and the air return inlets could circulate air between the nearby counters. It may cause unsterilized and unfiltered air circulating to nearby booths and increased the risk of exposure to viral aerosols.

Last, insufficient hygiene facilities may also increase the risk of transmission. At least 10,000 personnel visit the market per day, so is relatively inadequate that there was only one toilet for men and women. This further lead to gathering of waiting personnel in or outside the toilets. Infected personnel with symptoms, such as diarrhea, vomiting, cough or sneeze, may easily spread viral droplets or aerosols to the crowd or surroundings by high viral laden excretions. In toilet, infected personnel may simply spread the virus by walking around or touching object surfaces with contaminated soles or hands. With insufficient hand washing facilities making the situation even worse, contact transmission of the virus may occur more frequently and serious than expected.

Altogether, the initial Salmon-to-human transmission at Booth#14 in aquatic product area led to the viral spread in the Xinfadi Market 5 . The closer distance between aquatic product and soy product areas as well as the characteristics of higher density of population (e.g., sellers, buyers, and visitors), low temperature, high humidity, poor ventilation and hygiene, and lack of disinfection may contribute to the highly cluster cases in these two areas. This is not unique in Xinfadi Market since cluster epidemics have also been reported within other seafood markets (e.g., Huanan Seafood Wholesale Market in Wuhan), seafood industries 3 and J o u r n a l P r e -p r o o f meat processing facilities globally 4 . Therefore, wholesale markets with these characteristics may face outbreak risks.

Based on what we have found, precautionary control strategies that are important for public health protection are needed in wholesale markets during the pandemic of COVID-19

to limit future transmission of infectious viruses 34 with strict adherence to hand hygiene (e.g., hand washing, using alcohol hand rub or hand sanitizer) and social distance to avoid potential infection 35 . Hand sanitizers or quick drying hand disinfectants should be available for people entering the markets.

There are limitations in this study. First of all, we were not able to use actual virus or pseudovirus on-site, and the particle sizes of SARS-CoV-2 and fluorescent microspheres are different, which in this case may lead to certain bias from real conditions. Our results could 

The authors declare no conflict of interest.

The Supporting Information includes additional information about details of samples collected in aquatic, soy and toilet area and PPE, comparison between pseudovirus and fluorescent microsphere, and representative photo of air conditioner of the market. . Schematic diagrams of representative sampling points and positive points in the aquatic products area (A), soy products area (B), and the toilet (C). Smear samplings for several items (e.g., knife, calculator and platform scale button) are not shown. Each counter is numbered as A# and S# for aquatic products and soy products areas, respectively. For the toilet, each cubicle is numbered as T#. A window and hand washing facilities are also shown. Bright gray and dark gray indicate the counters/toilet/cubicle where the simulation experiment was conducted, respectively. Blue strips around the counters indicate the drainage channels. The points stuck on the side of the counters indicate that samples were smeared or collected between the countertop and the ground. 

Covid-19: WHO raises concerns about new cases in Beijing

Time Course of a Second Outbreak of COVID-19 in Beijing, China

COVID-19 Outbreak In Pacific Northwest Seafood Industry As Season Ramps Up

COVID-19 Among Workers in Meat and Poultry Processing Facilities -19 States

Cold-chain food contamination as the possible origin of COVID-19 resurgence in Beijing

Update: COVID-19 Among Workers in Meat and Poultry Processing Facilities -United States

Aerosol and surface contamination of SARS-CoV-2 observed in quarantine and isolation care

One Hundred Days of Coronavirus Disease 2019 Prevention and Control in China

COVID-19 Outbreak Associated with Air Conditioning in Restaurant

Aerosol and Surface Distribution of Severe Acute Respiratory Syndrome Coronavirus 2 in Hospital Wards

Transmission routes of respiratory viruses among humans

Aerosol transmission of SARS-CoV-2? Evidence, prevention and control

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

Detection of air and surface contamination by SARS-CoV-2 in hospital rooms of infected patients

Toilets dominate environmental detection of SARS-CoV-2 virus in a hospital

First Case of 2019 Novel Coronavirus in the United States

Clinical Characteristics of Coronavirus Disease 2019 in China

Comparison of different samples for 2019 novel coronavirus detection by nucleic acid amplification tests

Virological assessment of hospitalized patients with COVID-2019

Viable SARS-CoV-2 in various specimens from COVID-19 patients

Aerosol Generation by Modern Flush Toilets. Aerosol science and technology : the journal of the American Association for

Can a toilet promote virus transmission? From a fluid dynamics perspective

Aerodynamic analysis of SARS-CoV-2 in two Wuhan hospitals

Aerosol and Surface Stability of SARS-CoV-2 as Compared with SARS-CoV-1

First confirmed detection of SARS-CoV-2 in untreated wastewater in Australia: A proof of concept for the wastewater surveillance of COVID-19 in the community

Presence of SARS-Coronavirus-2 RNA in Sewage and Correlation with

Prevalence in the Early Stage of the Epidemic in The Netherlands

SARS-CoV-2 RNA in wastewater anticipated COVID-19 occurrence in a low prevalence area

SARS-CoV-2 in Detroit Wastewater

Stability of SARS-CoV-2 in different environmental conditions

Temperature-dependent surface stability of SARS-CoV-2

Likelihood of survival of coronavirus in a respiratory droplet deposited on a solid surface

Survival Characteristics of Airborne Human Coronavirus 229E

Effects of Air Temperature and Relative Humidity on Coronavirus Survival on Surfaces

Applying principles of behaviour change to reduce SARS-CoV-2 transmission

Mask use during COVID-19: A risk adjusted strategy

The authors gratefully acknowledge all CDC colleagues and healthcare workers involved in the epidemiological investigation, disinfection, and diagnosis and treatment of patients 

J o u r n a l P r e -p r o o f