Multi-route transmission potential of SARS-CoV-2 in healthcare facilities.

Understanding the transmission mechanism of SARS-CoV-2 is a prerequisite to effective control measures. To investigate the potential modes of SARS-CoV-2 transmission, 21 COVID-19 patients from 12-47 days after symptom onset were recruited. We monitored the release of SARS-CoV-2 from the patients' exhaled breath and systematically investigated environmental contamination of air, public surfaces, personal necessities, and the drainage system. SARS-CoV-2 RNA was detected in 0 of 9 exhaled breath samples, 2 of 8 exhaled breath condensate samples, 1 of 12 bedside air samples, 4 of 132 samples from private surfaces, 0 of 70 samples from frequently touched public surfaces in isolation rooms, and 7 of 23 feces-related air/surface/water samples. The maximum viral RNA concentrations were 1857 copies/m3 in the air, 38 copies/cm2 in sampled surfaces and 3092 copies/mL in sewage/wastewater samples. Our results suggest that nosocomial transmission of SARS-CoV-2 can occur via multiple routes. However, the low detection frequency and limited quantity of viral RNA from the breath and environmental specimens may be related to the reduced viral load of the COVID-19 patients on later days after symptom onset. These findings suggest that the transmission dynamics of SARS-CoV-2 differ from those of SARS-CoV in healthcare settings.

Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first identified from a cluster of patients with pneumonia of unknown cause in Wuhan, China, in December 2019. SARS-CoV-2 is an enveloped, positive-sense, single-stranded RNA virus and the seventh member of the coronavirus family that infects humans. It is distinct from both severe acute respiratory syndrome coronavirus (SARS-CoV) and Middle East respiratory syndrome coronavirus (MERS-CoV) (Zhu et al., 2020). Infection with SARS-CoV-2 may result in a wide range of clinical manifestations from asymptomatic infection to critical illness. As of August 11, 2020, COVID-19 has caused a worldwide pandemic, with nearly 20 million confirmed cases and 0.74 million deaths (World Health Organization, 2020a).

SARS-CoV-2 is believed to spread efficiently, but the exact transmission routes, especially the fomite, airborne (or aerosol), and fecal-oral routes, remain under debate (World Health Organization, 2020b). Viral RNA has been detected from various surfaces of isolation rooms and intensive care units in which COVID-19 patients are treated, but detection from air has been relatively difficult (Chia et al., 2020; Ong et al., 2020). The probable explanation is either that patients released limited virus during the sampling period or that patient’s exhaled breath (EB) is readily diluted by the room’s ventilation before it can be sampled, which is consistent with the proximity effect in the transmission of respiratory infections (Liu et al., 2017). Although SARS-CoV-2 may be transmitted by transmission routes similar to those of other coronaviruses, such as large droplets, aerosols, or contact, the potential of transmission via each route requires systematic investigation.

Nosocomial transmission of the virus has been reported in China: more than 3000 medical staff members have been infected during medical practice (Zhang, 2020). According to a single-center case series involving 138 hospitalized patients in Wuhan, China, nosocomial transmission was suspected to occur via 29 % of infected medical staff and 12.3 % of hospitalized patients (Wang et al., 2020). Similarly, the transmission of virus from a single health care worker to several other patients and health care workers was assumed to be responsible for the hospital outbreak of SARS-CoV-2 in the University Hospital of Münster, German (Schwierzeck et al., 2020). An accumulating body of evidence suggests that although COVID-19 patients are mostly infectious during the early stage of infection, the high viral loads in their sputum and feces specimens during later stages of infection may still pose a risk to healthcare workers (Zheng et al., 2020). This hypothesis has yet to be systematically evaluated.

In this study, we aimed to investigate the EB and the hospital environmental contamination by COVID-19 patients in the later stage of infection. To that end, we sampled the EB, exhaled breath condensate (EBC), surfaces of isolation wards, and personal necessities of 21 patients with laboratory-confirmed COVID-19 and hospital drainage systems from 12–47 days after symptom onset and subjected the samples to SARS-CoV-2 detection by quantitative reverse transcription polymerase chain reaction (RT-qPCR).

Methods

Patients

Twenty-one patients with laboratory-confirmed COVID-19 who were hospitalized in the First Affiliated Hospitals, College of Medicine, Zhejiang University, from February 13, 2020, to March 5, 2020, were enrolled in this study. We collected samples of EB, EBC, size-segregated aerosols from the air, frequently touched surfaces, and the drainage system. Patients’ sputum and feces specimens were collected daily by the hospital staff for purposes of routine diagnosis. The patients’ clinical data were obtained from hospital electronic medical records. This study protocol was approved by the Clinical Research Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University.

EB sampling

An exhaled aerosol collection system was developed (Fig. S1), and the patients were asked to breath normally through a mask for 30 min, during which they were asked to perform 10 forced coughs. The mask did not interrupt the patient’s oxygen inhalation. The exhaled air was sampled by the NIOSH bioaerosol sampler, which collected air at 3.5 L/min (105-L air was sampled during the 30-min sampling) and separated virus-laden aerosols into three size fractions: <1 μm, 1–4 μm, and >4 μm (Lindsley et al., 2010). After collection of EB samples, the NIOSH bioaerosol sampler was disassembled in a biosafety cabinet, and 1 mL of viral transport medium was added to each collection tube and PTFE filter to suspend virus particles.

EBC sampling

The EBC samples were collected using a sterile laboratory-made EBC collection system comprising a 15-mL centrifuge tube with the bottom cut off and a 50-mL centrifuge tube (Fig. S2). The patients were asked to blow into the 15-mL centrifuge tube for 10 min. Approximately 200–500 μL of EBC was collected from each patient for further analysis.

Room air and frequently touched surface sampling in isolation rooms

For sampling of isolation room air, a NIOSH sampler was placed on a tripod 1.2 m in height and 0.2 m away from the bed at the side of the patient’s head (Fig. S3). The sampling duration was 30 min, and a total of 105-L room air was sampled.

Frequently touched public and private surfaces in isolation rooms were sampled, as illustrated in Fig. S3 and Table S1. The standard sampling area was 10 × 10 cm2 (swab applied in horizontal followed by vertical and diagonal sweeps). For surfaces smaller than 10 × 10 cm2, the entire touchable surface was swabbed. After sampling, the swabs were immediately placed into 1.5 mL of viral transport medium.

Drainage-related sampling

Virus-laden aerosols may be generated during flushing in the toilet and in sewage pipes, accounting for potential fecal-oral transmission of SARS-CoV-2. For lavatory bioaerosol sampling, a NIOSH sampler was placed on a tripod 1.2 m in height and 0.5 m away from the toilet bowl in the lavatory of the isolation rooms; the 30-min sampling period began when the patient used the toilet for defecation. After the patient exited the toilet, the toilet bowl and floor drain were also swabbed for sampling. In addition, water samples were taken in 15-mL centrifuge tubes from the main sewage pipe and wastewater pipe of the building in which the COVID-19 patients were isolated (Fig. S4).

RT-qPCR

Viral RNA was extracted from all collected clinical or environmental specimens using MagNA Pure LC 2.0 (Roche, Basel, Switzerland). RT-qPCR was performed using a China Food and Drug Administration-approved commercial kit specific for SARS-CoV-2 detection (Da’an Gene Co., Ltd., Guangzhou, Guangdong, China) according to the manufacturer’s protocol. Specimens with quantification cycle (Cq) values of 40 or lower were considered positive for SARS-CoV-2 RNA. Specimens with Cq values higher than 40 were tested again, and those with repeat Cq values that were higher than 40 or undetectable were considered negative for SARS-CoV-2 RNA. The standard curve was generated by 10-fold serial dilutions of positive control and the fitted relationship between Cq value and virus concentration in the tested medium, C (copies/mL), is as follows

Results and discussions

Detection of SARS-CoV-2 RNA among various types of samples

A total of 21 moderately to critically ill patients with COVID-19 (aged 13–72 years; median, 61 years) were recruited. Sampling was performed 12–47 DAO (median, 29 days). The patients’ details are listed in Table 1 . All recruited patients were in the recovery period with normal body temperature and improved lung infection conditions according to computed tomography scan. Viral loads in the sputum specimens were detected varying from negative to 1.2 × 1010 copies/mL (median, 1.3 × 105), and those in feces specimens from negative to 1.9 × 109 (median, 7.1 × 103).

Table 1

Information of COVID-19 patients and sampling types.

Patient No.	Age (y)	Sex	Disease severity	Sampling type	Time of sample collection (DAO)	Viral load of sputum/feces specimen (log10 copies/mL)	Body temperature (°C)
1	66	Male	Critical	EB	14	5.2/3.9	36.8
2	48	Female	Severe	EB	13	4.6/3.2	36.6
3	64	Male	Critical	EB	18	6.9/9.3	36.5
4	67	Female	Severe	EB	21	3.5/ND	37
5	54	Male	Severe	EB	20	6.5/4.7	36.5
				Room air, surfaces*	37	5.0/ND	36.6
				EBC	41	6.3/ND	36.6
6	61	Male	Critical	EB	15	7.5/4.2	36.6
7	36	Male	Moderate	EB, Lavatory	21	10.1/5.6	36.9
8	30	Male	Moderate	EB, Lavatory*	22	6.1/5.3	37.2
9	62	Male	Severe	EB, Lavatory	28	7.0/5.7	37.6
				Room air*, surfaces	43	5.5/3.6	36.9
				EBC*	47	5.8/2.8	36.9
10	62	Male	Severe	Lavatory*	12	Neg/6.2	36.6
11	13	Male	Moderate	Lavatory	18	Neg/6.0	36.7
12	51	Male	Critical	Lavatory	24	Neg/6.7	36.9
				Room air, surfaces	38	5.4/ND	36.4
				EBC	42	Neg/Neg	37.1
13	70	Male	Critical	Room air, surfaces	31	4.2/4.6	37.1
				EBC	35	ND/5.3	36.8
14	61	Male	Severe	Room air, surfaces	29	5.7/Neg	36.5
				EBC	33	5.2/ND	36.6
15	38	Female	Moderate	Room air, surfaces*	29	5.1/Neg	36.9
				EBC	34	2.8/ND	36.8
16	62	Female	Critical	Room air, surfaces*	23	Neg/Neg	37.2
				EBC	27	Neg/Neg	37.1
17	70	Male	Severe	Room air, surfaces	29	4.6/ND	36.9
				EBC*	34	6.2/Neg	37.0
18	57	Female	Moderate	Room air, surfaces*	39	Neg/3.9	37.2
19	29	Female	Severe	Room air, surfaces	31	Neg/Neg	37.2
20	72	Male	Critical	Room air, surfaces	35	Neg/Neg	36
21	66	Male	Critical	Room air, surfaces	31	Neg/Neg	37

Notes:

1. Positive samples are highlighted with asterisk (*).

2. ND is short for “not delivered”, which means the sputum or feces specimen was not sampled.

3. Neg is short for negative.

Of the 254 samples (9 EB samples, 8 EBC samples, 12 bedside air samples, 202 public/private surface samples from isolation rooms, and 23 feces-related air/surface/water samples), 14 tested positive, as summarized in Table 2 , suggesting the potential for nosocomial transmission of SARS-CoV-2 via multiple routes. The viral RNA concentrations were 1112 copies/m3 and 745 copies/m3 respectively in the <1 μm and >4 μm fractions of the only positive air sample. The maximum viral RNA concentrations were 38 copies/cm2 in sampled private surfaces of COVID patients (excluding the toothbrush) and 3092 copies/mL in sewage/wastewater samples. These findings prioritize areas for effective environmental disinfection practices to reduce virus transmission and support the need for strict adherence to personal hygiene. However, the positive detection rates and virus contents in these positive samples were relatively low, indicating that patients release limited virus particles into the environment more than 10 DAO.

Table 2

Summary of the sampling results.

Sampling type	Time of sample collection (DAO)	Viral load of sputum/feces specimen (log10 copies/mL)#	Positive rates	Viral RNA concentration in positive samples
Exhaled breath	13–28 (median, 20)	3.5–10.1/5.0–9.3	0/9	/
Exhaled breath condensate	27–43 (median, 30)	Neg&–6.3/ Neg–5.3	2/8	216 copies/mL, Patient 9
				222 copies/mL, Patient 17
Room air at patient’s bedside	23–43 (median, 31)	Neg–5.7/ Neg–4.6	1/12	1112 copies/m3 (<1 μm) and 745 copies/m3 (>4 μm), Patient 9
Private surfaces in isolation rooms			4/132	9 copies/cm2, towel, Patient 5
				1045 copies in total from the toothbrush of Patient 15*
				35 copies/cm2, towel, Patient 16
				38 copies/cm2, bedside wall, Patient 18
Public surfaces in isolation rooms			0/70	/
Drainage systems (lavatory air, toilet bowl, floor drain)	12–28 (median, 21.5)	6.1–10.1/5.3–6.7	0/6, 2/6, 1/6	4 copies/cm2, toilet bowl, Patient 8
				2 copies/cm2, toilet bowl, Patient 10
				2 copies/cm2, floor drain, Patient 10
Drainage systems (Sewage, wastewater)	/	/	4/5	589 copies/mL, inlet of the septic tank
				3092 copies/mL, outlet of the septic tank
				1660 copies/mL, inlet of the wastewater pipe
				363 copies/mL, inlet of sewage treatment station

Sputum or feces specimens of the patients were occasionally not delivered for PCR analysis on the sampling date and were not considered in the statistics.

10 mL water was used to rinse the toothbrush, and the water was directly analyzed by the PCR kit.

Neg is short for negative.

Potential of respiratory bioaerosols in transmitting SARS-CoV-2

Respiratory droplets, or bioaerosols, are carriers of pathogens and are responsible for the transmission of respiratory infectious diseases. They are produced during expiratory activities such as coughing, sneezing, talking, and normal breathing. To test the potential of respiratory bioaerosol in transmitting SARS-CoV-2, 9 EB and 8 EBC specimens were collected from 15 patients between 13 and 43 DAO in the present study. The aerosol collection system in this study directly sampled patients’ EB before it was highly diluted by the room air, but all samples in three size fractions tested negative for viral RNA [positive detection in sputum samples; see Table 2], though 10 forced coughs were performed by each patient. Two of the eight EBC samples tested positive for the virus (Patients 9 and 17), with the virus concentration as 216.0 copies/mL and 222.0 copies/mL, respectively. This detection rate was likely attributable to the high collection efficiency of both fine and coarse aerosols and the low possibility of dilution by air or viral transport medium. Furthermore, 1 (Patient 9) of the 12 air samples from the patients’ bedsides (24–43 DAO) tested positive (sputum viral loads between negative and 4.5 × 105 copies/mL); RNA was detected in two size fractions from that air sample, with virus concentrations of 1111.9 copies/m3 and 744.6 copies/m3 in the <1 μm and >4 μm fractions, respectively. One of the two positive EBC samples and the positive bedside air sample belonged to Patient 9; his sputum viral load values on the sampling date were 6.3 × 105 copies/mL and 3.2 × 105 copies/mL, respectively. These findings are consistent with a sampling event in two Wuhan hospitals where viral RNA in aerosols detected in isolation wards and ventilated rooms was very low (Liu et al., 2020). In another sampling event with NIOSH samplers, viral RNA was readily detected from the air of isolation rooms where patients were at 5 D.A.O.; the highest concentration was 1384 copies/m3 and 2000 copies/m3 in the 1−4 μm and >4 μm fractions, respectively, and no viral RNA was detected in <1 μm size fraction (Chia et al., 2020).

Although some COVID-19 patients evaluated during the later stage of infection in our study had a relatively high viral load in their sputum specimen, shedding of virus particles from the respiratory tract into the environment was limited, which is in agreement with previous findings that viral load in nasal or throat samples is high in the early stage of infection and decreases to a low level at around 9 DAO (Woelfel et al., 2020; Zou et al., 2020). Virus-laden respiratory droplets from the deep lung tend to deposit and accumulate into the respiratory mucous, but difficult to follow the respiratory airflow to enter the indoor environment (Kleinstreuer and Zhang, 2010). The transmission characteristics of SARS-CoV-2 differ from those of SARS because the viral loads in the respiratory specimens of SARS-infected patients peak at 10 DAO (Peiris et al., 2003), and considerable positive rates of surface swabs have been detected in SARS patients’ rooms at 5–15 DAO (Dowell et al., 2004). However, the virus-shedding course of SARS-CoV-2 is similar to that of influenza virus (Tsang et al., 2015; Zou et al., 2020), and the NIOSH bioaerosol sampler has also successfully detected influenza virus in samples of 32 out of 38 influenza-positive patients, 60 % of whom had symptom onset within 3 days (Lindsley et al., 2010). This finding highlights the importance of early control measures for COVID-19 patients and suggests a relief for medical staff working in isolation rooms and intensive care units with recovering (later stage) patients.

Personal necessities function as potential vehicles for SARS-CoV-2 transmission

Surfaces may be contaminated by deposition of virus-laden droplets or by contact of patients’ hands, and lead to fomite transmission of SARS-CoV-2. Four of the 202 surface samples obtained from seven isolation rooms tested positive. These positive samples were from private surfaces of COVID-19 patients, namely the towels of Patient 5 (9 copies/cm2) and Patient 16 (35 copies/cm2), the toothbrush of Patient 15 (1045 viral RNA copies in total), and the bedside wall of Patient 18 (38 copies/cm2) (Table 2). This positive detection can be attributed to contamination with respiratory secretions. It is worth noting that although both sputum specimen of Patient 16 and 18 tested negative, viral RNA was still detected from their towel or room (the bedside wall), which might be shed days ago. Viruses may survive on surfaces for hours to days and might infect susceptible individuals (van Doremalen et al., 2020). The viral survival, the transmission efficiency of the virus from surfaces and hands to respiratory mucosa, and the dose-response rate after the virus is introduced onto the mucosa are all important factors that influence the exposure risk of the susceptible individuals. Our results are in line with that of a previous study that revealed that towels in hospitals are frequently contaminated with respiratory pathogens (Hassan et al., 2019). These findings suggest that personal necessities function as potential vehicles for virus transmission. Therefore, measures that target the proper use of personal necessities, including avoidance of sharing with others, frequent cleaning with disinfectant, and hand hygiene practices after use, may help to limit virus transmission.

Studies have reported much high positive rates from public surface samplings near COVID-19 patients within 5 DAO (Chia et al., 2020; Ong et al., 2020). However, no viral RNA was detected from samples taken from frequently touched public surfaces in the isolation rooms over 7 h after routine sanitization, further indicating limited transmission potential via the fomite route during our sampling period. In previous studies, viral RNA were readily detected from public surfaces in isolation rooms of patients in the early stage of infection (Chia et al., 2020; Ong et al., 2020). The viral loads of the sputum and feces specimen of 12 patients from the isolation rooms (Table S1) in our study were negative–4.5 × 105 copies/mL and negative–3.7 × 104 copies/mL, respectively, on the sampling date.

Potential of fecal-oral and fecal-aerosol routes in transmitting SARS-CoV-2

COVID-19 patients have a long virus shedding course in feces specimen, and viral RNA, even viable virus particles, has been detected in the feces-related specimens of patients with confirmed COVID-19 (Guan et al., 2020; Tian et al., 2020). Moreover, previous studies have suggested that the viral loads of feces specimens peak late after the symptom onset and may last for a longer duration compared with those of respiratory specimens (Xing et al., 2020). Thus, fecal-oral and fecal-aerosol transmission have been suspected as potential routes of SARS-CoV-2 (McDermott et al., 2020; Tian et al., 2020).

We sampled lavatory air, toilet bowl, floor drain and sewage, apart from the public lavatory surfaces as described above and in Fig. S3. Two of the six samples from toilet bowl after defecation and toilet flushing tested positive for the viral RNA, namely 4 copies/cm2 and 2 copies/cm2, respectively (viral loads for positive detection in the feces specimens, 2.1 × 105 copies/mL–5.0 × 106 copies/mL). Further, one of six floor drain samples tested positive, with a viral concentration of 2 copies/cm2, indicating that virus-laden aerosol transmission by the drainage system (as illustrated in Graphical Abstract) is probable. None of the six lavatory aerosol samples tested positive; although patients’ feces specimen were test high in viral loads, the toilet didn’t produce detectable amount of virus-laden aerosols in our sampling event. Moreover, four of the five water samples from the drainage system tested positive. The viral RNA concentration of the sewage is as high as 3092 copies/mL, but the sewage treatment station worked well to reduce it below the detection limit (Fig. S4).

The detection of SARS-CoV-2 in the toilet bowl and sewage samples in this study is consistent with that finding, indicating high viral loads in the patients’ feces-related specimens. Notably, one sample from the floor drainage tested positive, which was probably contaminated by the virus-laden bioaerosols from the drainage system. The toilet flushing process produces large numbers of bioaerosols in the sewage pipe, which may re-enter vertically-aligned lavatories in the same building through the floor drain. This likely explains the SARS cluster outbreak that occurred in Amoy Garden in 2003 (Yu et al., 2004).

Implications

Our findings provide important health implications for the control of nosocomial infection and personal protection: (1) ensure good ventilation for each isolation ward to reduce the potential risk of transmission via aerosols; (2) in addition to routine disinfection practices, perform frequent and thorough environmental cleaning and disinfection of areas not easy to clean; (3) perform ventilation and disinfection after using the toilet to reduce aerosolization of virus and contamination of surrounding surfaces, and preserve tap water seals to prevent the re-entry of bioaerosols from pipes; (4) do not share personal necessities, and disinfect and replace them regularly. However, the major limitation is that we only detected the presence of the nucleic acid (RNA) of SARS-CoV-2 and did not perform viral culture due to the low virus quantity in the positive samples. The presence of nucleic acid does not confirm the existence of viable virus, nor does it indicate that fomite contact is an important route for SARS-CoV-2 transmission.

CRediT authorship contribution statement

Baihuan Feng: Conceptualization, Data curation, Methodology, Writing - original draft, Writing - review & editing. Kaijin Xu: Data curation, Methodology. Silan Gu: . Shufa Zheng: Data curation, Methodology. Qianda Zou: Data curation, Validation. Yan Xu: Data curation, Validation. Ling Yu: . Fangyuan Lou: Data curation, Validation. Fei Yu: Data curation, Validation. Tao Jin: Project administration. Yuguo Li: Project administration. Jifang Sheng: Project administration. Hui-Ling Yen: Project administration. Zifeng Zhong: Project administration. Jianjian Wei: Funding acquisition, Project administration, Supervision, Writing - original draft, Writing - review & editing. Yu Chen: Project administration, Supervision, Writing - original draft, Writing - review & editing.

Declaration of Competing Interest

The authors report no declarations of interest.