SARS-CoV-2 and the role of orofecal transmission: a systematic review

Background: Modes of transmission of SARS-CoV-2 are of key public health importance. SARS-CoV-2 has been detected in the feces of some COVID-19 patients, suggesting the possibility that the virus could, in addition to droplet and fomite transmission, be transmitted via the orofecal route. Methods: This review is part of an Open Evidence Review on Transmission Dynamics of COVID-19. We conduct ongoing searches using WHO COVID-19 Database, LitCovid, medRxiv, and Google Scholar; assess study quality based on five criteria and report important findings. Where necessary, authors are contacted for further details on the content of their articles. Results: We include searches up until 20 December 2020. We included 110 relevant studies: 76 primary observational studies or reports, and 35 reviews (one cohort study also included a review) examining the potential role of orofecal transmission of SARS-CoV-2. Of the observational studies, 37 were done in China. A total of 48 studies (n=9,081 patients) reported single cases, case series or cohort data on individuals with COVID-19 diagnosis or their contacts and 46 (96%) detected binary RT-PCR with 535 out of 1358 samples positive for SARS-CoV-2 (average 39.4%). The results suggest a long duration of fecal shedding, often recorded after respiratory samples tested negative, and symptoms of gastrointestinal disease were reported in several studies. Twenty-nine studies reported finding SARS-CoV-2 RNA in wastewater, river water or toilet areas. Six studies attempted viral culture from COVID-19 patients’ fecal samples: culture was successful in 3 of 6 studies, and one study demonstrated invasion of the virus into intestinal epithelial cells. Conclusions: Varied observational and mechanistic evidence suggests SARS-CoV-2 can infect and be shed from the gastrointestinal tract, including some data demonstrating viral culture in fecal samples. To fully assess these risks, quantitative data on infectious virus in these settings and infectious dose are needed.


Introduction
Understanding how, when and in what types of settings SARS-CoV-2 spreads between people is critical to developing effective public health and infection prevention and control measures to break the chains of transmission. Current evidence suggests SARS-CoV-2 is primarily transmitted via respiratory droplets and fomites between infected individuals and others in close contact 1 .
SARS-CoV-2 has been shown to contaminate and survive on certain surfaces, and has also been detected in the feces of some patients, suggesting the possibility that SARS-CoV-2 could transmit via the orofecal route, including potentially via contamination of food. It is well recognized that coronaviruses are major pathogens in many mammalian species and predominantly target epithelial lining cells in the respiratory and gastrointestinal (GI) tracts. Many animal coronaviruses are transmitted by the fecal-oral route and there are several reports of intestinal disease associated with SARS-CoV-1 and other human coronaviruses. Main causes include lack of adequate sanitation and poor hygienic practices. Identifying infectious virus and quantifying viral load within human GI tissues, feces, and contaminated materials including fomites and sewage would help understand the potential for transmission. We aimed to systematically review the evidence on orofecal SARS-CoV-2 transmission. Terminology for this article can be found in Box 1.

Methods
We are undertaking an open evidence review investigating factors and circumstances that impact on the transmission of SARS-CoV-2, based on our protocol (see Extended data: Appendix 1 2 ). For the original protocol, see https://www. cebm.net/evidence-synthesis/transmission-dynamics-of-covid-19/. In brief, this review aims to identify and evaluate relevant articles (peer-reviewed or awaiting peer review) that examine the mode of viral transmission and ecological variables influencing the mode of transmission. We conduct an ongoing search using WHO COVID-19 Database, LitCovid, medRxiv and Google Scholar for keywords and associated synonyms. Results are reviewed for relevance and for articles that looked particularly relevant forward citation matching was undertaken and relevant results were identified. Studies with modelling are only included if they report transmission outcome data and not predicted outcomes (see further details of the search strategy in the Extended data: Appendix 2 2 ). Searches are updated every two weeks.
We extracted data on the type of study, setting, sample source and methods, fecal PCR positive samples for SARS-CoV-2 RNA including cycle threshold (including methods), symptom chronology in relation to PCR testing and/or taking samples and viral culture. We tabulated the data and summarised data narratively by mode of sample. We assessed quality using a modified QUADAS-2 risk of bias tool 3 . We simplified the tool as the included studies were not designed as primary diagnostic accuracy studies and assessed study quality based on five criteria. Where necessary we write to authors of included studies for further details or clarification on the content of their articles. Meta-analyses were not performed, due to the variability of available data. The protocol was last updated on 1 December 2020 (Version 3: 1 December 2020).

Results
This update includes searches up until 20 December 2020 (see Figure 1). We identified 110 relevant studies (see Extended data: Appendix 3 for references of included studies 2 ): 76 primary studies or reports, and 35 reviews, one of which also reported primary study results from a cohort study [Cheung K 2020].

Reviews
The included reviews summarised a range of observational studies including studies of detection of SARS-CoV-2 RNA in fecal samples of individuals testing positive for SARS-CoV-2 in respiratory samples, frequency of GI symptoms among those with COVID-19, and observations of SARS-CoV-2 RNA in toilets and wastewaters. The reviews included overlapping studies and must therefore not be considered as entirely additional information. Five followed systematic review methodology and reporting [Edwards 2020, Karia 2020, Pamplona 2020, Parasa 2020, Santos 2020]. The quality of the other reviews was low to moderate, with none assessing included study quality, and with reporting of methods often missing or very limited.
None of these reviews focussed on the infectiousness (and hence transmission potential) of SARS-CoV-2 identified in fecal

Amendments from Version 1
Thank you to both reviewers who offered suggestions to improve the manuscript; we have made additions and edits to include their suggestions. We have added a definition for viral load. We have added an explanation that these studies do not provide an estimate of infectious virus concentration and that it is not reasonable to deduce human transmission risk from PCR positive samples. We note that there may be overlap in included studies, but that we have no adequate means to identify this; we also explain that the reviews may also overlap. We note that limited data on sex or by age-group precluded additional investigation into these factors. We are grateful that the error in the table was noticed and we have deleted this unnecessary total number from the table. The issue of how many studies were available as preprints and how these varied by study type is interesting but beyond the scope of this review, not least because this changes over time. The next time we update this review we hope to have a better breakdown of those fully published, and hopefully a better bank of evidence to analyse. or wastewater samples. A review on the potential for foodborne transmission of SARS-CoV-2 found no published studies of SARS-CoV-2 survival in or on food products. The totality of the reviews' evidence shows that the SARS-CoV-2 RNA is commonly present in stool samples of COVID-19 patients but it is unknown if this represents primary invasion of enterocytes or simply saliva and sputum that has been swallowed and is transiting it way through the GI tract. The presence of viral RNA in the feces does not allow any conclusions to be drawn about infectiousness. The contribution of orofecal transmission to viral spread in the pandemic has not been established or quantified.

Primary studies
Quality of included studies. Overall the quality of the evidence was low to moderate mainly due to a lack of standardisation of techniques, omissions in reporting and a failure to account for biases in the research process (see Table 3; Figure 2). Sample sources were clear in two-thirds of studies (65.8%). Several studies mention the possibility of bias influencing their findings but did not use strategies (design or analysis) to deal with bias, and as such have been recorded as unclear risk of bias.

Results.
A total of 48 (n=9,081 patients) studies reported single cases, case series or cohort data on individuals with COVID-19 diagnosis or their contacts and 46 studies (96%) detected binary RT-PCR with 535 out of 1358 samples RT-PCR positive for SARS-CoV-2 (average 39.4%). All but five studies were in hospitalized patients; 31 were done in China; the others were in East Asia, South East Asia, South Asia, USA and Europe (see Table 1).

PCR testing for SARS-CoV-2 RNA in fecal samples in hospitalised COVID-19 patients.
Of the 43 hospital studies, 41 (95%) detected binary RT-PCR, with 522 positive tests out of 1293 fecal samples (average 40.4%) from COVID-19 patients (see Table 4). These studies were mainly small case series, they included patients of a range of ages from infancy to elderly and with widely varying severity of disease, and the proportion of fecal samples varied from 1 (nine studies) to 258 [Zhang Y, Chen C], and the proportion testing positive for SARS-CoV-2 RNA varied from 14% to 100% across studies. One study that identified SARS-CoV-2 RNA in fecal samples among 39 of 73 hospitalised COVID-19 patients, also studied the gastric, duodenal and rectal epithelia of one patient using specimens collected via endoscopy [Xiao F, Tang M 2020].
Immunofluorescence data showed that ACE2 protein, proven to be a cell receptor for SARS-CoV-2, was abundantly expressed in the glandular cells of gastric, duodenal, and rectal epithelia, supporting entry of SARS-CoV-2 into the host cells. Intracellular staining of viral nucleocapsid protein in gastric, duodenal, and rectal epithelium showed that SARS-CoV-2 infects these GI glandular epithelial cells. Viral RNA was also detected in esophageal mucous tissue, but a lack of viral nucleocapsid protein staining in esophageal mucosa suggested low viral infection there. Viral nucleocapsid protein in rectal epithelial cells was detected in specimens from some additional COVID-19 patients, suggesting that infectious SARS-CoV-2 can survive the GI environment.

GI symptoms among COVID-19 patients
Reporting of GI symptoms among COVID-19 patients is frequent but not consistent within these studies (diagnosis is typically based on fever, respiratory symptoms and the results of PCR testing in respiratory swabs, so recording GI symptoms may not be routine). However, several observational studies report the presence of GI symptoms among COVID-19, including Chan 2020, Cheung 2020 and Han C 2020. GI symptoms do not necessarily correlate in severity or time with other COVID-19 disease symptoms.
Timing and duration of fecal shedding Fecal shedding of SARS-CoV-2 has been reported throughout the disease course and also continuing after respiratory samples tested negative. A five-person family with a confirmed COVID-19 case was hospitalized and observed: the parents and two children aged two and five years became infected but the youngest child was not infected. These children shed infectious virus via the respiratory system, and this shedding observed in the nasopharynx cleared after five to 6 days; however, viral RNA was continuously detected in the children's stool for more than four weeks [Wolf 2020]. Tang 2020 et al. reported an apparently asymptomatic (no fever or cough) 10-year-old child, from whom, 17 days after the last close contact with individuals testing positive for SARS-CoV-2, a fecal sample was positive for SARS-CoV-2 RNA. A retrospective study of 133 hospitalised COVID-19 patients identified 22 whose sputum or fecal samples tested positive after pharyngeal swabs became Additional evidence of SARS-CoV-2 replication activity was observed within the intestine: an inpatient for treatment of a rectal adenocarcinoma had samples taken from enteric sections, and the mucosa of rectum and ileum analysed [Qian 2020]. The rectal swab sample tested negative for SARS-CoV-2 RNA by PCR. However, typical coronavirus virions in rectal tissue were observed under electron microscopy with abundant lymphocytes and macrophages (some SARS-CoV-2 positive) infiltrating the lamina propria.
Methodological issues across these studies including variability in sample selection and methods of viral culture, reported in Table 5, mean the results may not be comparable and should be interpreted with caution.
All reported the detection of SARS-CoV-2 RNA and/or SARS-CoV-2 viral proteins; a number of studies suggested the potential of detection in sewage to be used as a public health monitoring system. River water. One study tested river water for SARS-CoV-2: taking samples from three urban river locations in a low sanitation urban context (i.e. highly impacted by raw sewage) in Quito, Ecuador, during a peak of COVID-19 cases. SARS-CoV-2 RNA was detected in all samples, at levels similar to those in wastewater from cities during outbreaks [Guerreo-Latorre 2020]. Evidence from reviews can be found in Table 2.

Discussion
The evidence from 110 relevant studies supports a potential role of orofecal transmission of SARS-CoV-2. Fecal shedding of SARS-CoV-2 RNA has been reported in 96% of the included observational studies, often for relatively long durations. Three studies reported the culture of SARS-CoV-2 using fecal samples, but requisite methods to confirm viral growth were lacking. One study demonstrated viral isolation from rectal tissue of a COVID-19 patient. Studies in hospitals show the presence of SARS-CoV-2 RNA at and around toilets and toilet rooms; there is evidence that disinfectant cleaning leaves no SARS-CoV-2 RNA detectable. Many studies report identifying SARS-CoV-2 RNA in sewage and wastewaters, but viral culture from such sources has not been demonstrated, so there is no evidence of infection risk from those sources; however, detection of SARS-CoV-2 RNA in sewage and wastewaters can be useful as a surveillance tool.
Experimental models of the human intestinal epithelium show that SARS-CoV-2 can infect this tissue and replicate, supporting the rationale for the human GI tract as a possible transmission route [4][5][6] ). Zang 2020 demonstrated that human enterocytes express high ACE2 receptor levels, supporting viral invasion at these sites 4 . Zang 2020 and Lamers 2020 et al. showed that SARS-CoV-2 productively infected human small intestinal organoids 4,5 . Zhou J 2020 et al. and co-workers showed active replication of SARS-CoV-2 in human intestinal organoids, and isolated infectious virus from the stool specimen of a patient with diarrheal COVID-19 6 . Yao et al. investigated the mutation spectrum, replication dynamics, and infectivity of 11 patient-derived SARS-CoV-2 isolates in diverse cell lines; the authors report that "three of our viral isolates were extracted from stool samples (two of which were very potent) indicating that viable SARS-CoV-2 particles could be found in stool samples" 7 .
These studies do not provide an estimate of infectious virus concentration; nor has the infectious dose for humans been established, so at this point it is unwarranted to deduce a human transmission risk based on this small number of virus-positive fecal samples. NR SARS-CoV-2 RNA is identified in a range of water environments including hospital wastewaters.
According to a few studies investigating the deactivation of SARS-Co V-2 showed that chlorinebased disinfectants are widely used for their broad sterilization spectrum, high inactivation efficiency and easy decomposition with little residue, as well as represents the best economic solution. The complete deactivation of SARS-CoV-2 can be achieved by combination of other technologies (biological and/or physical-chemical processes).
Authors suggest there is a need to develop more secure, efficient, economical disinfection technologies in order to limit

(NA)
Literature on SARS-CoV-2 in wastewaters is currently limited. For SARS-CoV-1, its resistance in wastewater is limited, especially at temperatures above 20 °C, and the virus has been easily removed with chlorine (> 0.5 mg L-1 for 30 min).
Detection of SARS-CoV-2 in wastewater might track the epidemic trends: although promising, an effective and wide application of this approach requires a deeper knowledge of the amounts of viruses excreted in faeces, and the actual detectability of viral RNA in sewage.
Cuicchi D 2020 no To collect the data available on SARS-CoV-2 in the GI system and evaluate whether the digestive system could contribute to viral transmission (31  The presence of SARS-CoV-2 RNA in river water and untreated wastewater is confirmed, but strong evidence of its survival time in water environments is missing. One study confirmed lack of infectivity of SARS-CoV-2 in water based on absence of cytopathic effect. There are more children than adults with asymptomatic infections, milder conditions, faster recovery, and a better prognosis. Some concealed morbidity characteristics also bring difficulties to the early identification, prevention and control of COVID-19.  Table 4.

Main findings of primary studies on orofecal transmission of SARS-CoV-2
Chan 2020 This very early study established the likelihood of person to person transmission of SARS-CoV-2, in hospital and family settings. The two faecal samples from patients 3 and 4 who had preceding diarrhoea were negative on a multiplex PCR assay for common diarrhoeal viruses, bacteria, and parasites.
Chen C 2020 This retrospective study of 133 hospitalised COVID-19 patients identified 22 whose sputum or fecal samples tested positive, after their pharyngeal swabs became negative.
Chen Y 2020 Sixty seven percent (28/42) laboratory-confirmed hospitalised COVID-19 patients tested positive for SARS-CoV-2 RNA in stool specimens; this was not associated with the presence of GI symptoms or severity of illness. Among them, 18 (64%) patients remained positive for viral RNA in the feces after the pharyngeal swabs turned negative, for a duration of 6 to 10 days.
Cheung CCL 2020 This study used multiplex immunohistochemistry and unexpectedly detected SARS-CoV-2 viral antigens in intestinal and liver tissues, in surgical samples obtained from two hospitalized patients who recovered from Covid-19. The presence of the virus was validated by RT-PCR and flow cytometry to detect SARS-CoV-2-specific immunity in the tissues.
Cheung K 2020 This study analysed stool samples from a cohort of 59 patients with COVID-19 in Hong Kong during February 2020 and additionally did a meta-analysis of data from 11 studies on the prevalence of GI symptoms and stool excretion of viruses. Fecal discharge continues long after respiratory shedding of COVID-19 has ceased.
Cho 2020 This case study reports an infant with mild Covid-19, positive-to-negative nasal swab conversion occurred on the 21st day from the onset of symptoms, but stool swab positivity persisted during the 6-week admission period and for 7 weeks during follow-up at an outpatient clinic after discharge.
Chu H 2020 Case report of a breastfeeding woman with a positive PCR test for SARS-CoV-2. The patient presented on 24 January 2020 with GI symptoms; later she developed a fever. Her infant had been born 16 January 2020. She tested PCR positive in respiratory swabs. She had persistent SARS-CoV-2 RNA positivity in her feces but negativity in her breastmilk. She bottle-fed her baby with her breastmilk after treatment. The baby appeared healthy and unaffected after a 1-month follow up.
COVID Ix Team SARS-CoV-2 RNA was detected in at least one nasopharyngeal (NP) swab, 11/12 oropharyngeal (OP) swab and 7/10 in the stool in this case series describing the first 12 US patients confirmed to have COVID-19 from 20 January to 5 February 2020.
Ge 2020 This case study of a hospitalised Covid-19 patient reported that the fecal samples remained PCR-positive for 22 days after their respiratory samples turned negative.
Han C 2020 Among a group of hospitalised patients with low severity COVID-19, digestive symptoms were present in 57%. Patients with digestive symptoms were more likely to be fecal virus-positive than those with respiratory symptoms.

Hayee 2020
This study reports PCR test results for outpatients attending for GI endoscopy at a UK hospital 30th April to 30th June 2020: 3/2,611 asymptomatic patients tested positive for SARS-CoV-2 on nasopharyngeal swab testing pre-endoscopy. No cases of Covid-19 were detected for 14 days after the procedure.
Hoehl S 2020 Children and staff at 50 day-care centres in Germany were tested repeatedly over 12 weeks. Buccal mucosal swabs and anal swabs were taken (by parents) from 825 children aged 3 months to 8 years attending the day care centres and 372 staff members (swabs self-collected) of these settings, between 18 June and 10 September 2020. 7,366 buccal mucosa swabs and 5,907 anal swabs were analysed. No respiratory or GI shedding of SARS-CoV-2 was detected in any of the children.
Two adult staff members at two different day care centers tested positive; one had symptoms.
Holshue 2020 Stool obtained from a single hospitalized Covid-19 case was positive for SARs-CO-V-2 on day 7 of the illness.
Jeong 2020 There was viable SARS-CoV-2 in saliva, urine, and stool from COVID-19 patients up until days 11 to 15 of the clinical course suggesting that viable SARS-CoV-2 can be secreted in various clinical samples as well as respiratory specimens.

Jiehao 2020
Prolonged virus shedding was observed in the respiratory tract and feces of children at the convalescent stage.

Main findings of primary studies on orofecal transmission of SARS-CoV-2
Kim J-M 2020 SARS-CoV-2 RNA was detected in serum, urine or stool samples in 20% of patients hospitalised with Covid-19: 13/129 stool samples obtained from 74 patients tested positive. However, the virus could not be isolated from these samples and therefore the risk of transmission via these media is not established.
Lescure 2020 A case series of the first five identified Covid-19 cases in Europe; an early demonstration of the vastly increased risk for elderly versus younger people. Viral RNA was detected in the stools of the two paucisymptomatic women.
Li 2020 This case series reported on 29 hospitalised mild-to-moderate severity Covid-19 patients in China. Fecal samples from 4 patients tested positive for SARS-CoV-2 RNA, including one apparently asymptomatic case.
Ling 2020 From 292 confirmed cases with COVID-19 in the Shanghai region, 66 recovered patients were included. Clearance of viral RNA in patients' stools was delayed compared to oropharyngeal swabs.
Lo 2020 A report of the clinical and microbiological features of ten hospitalized Covid-19 patients in Brazil between 21 January and 16 February 2020 found that SARS-CoV-2 can be shed in the stool.
Nicastri 2020 SARS-CoV-2 RNA was positive in stools, nasopharyngeal and oropharyngeal swabs at different time points in a case report.
Pan 2020 Stool samples from 9/17 confirmed patients were positive on RT-PCR analysis.
Peng 2020 Virus was found in urine, blood and in two anal swabs and oropharyngeal swabs of nine patients diagnosed with COVID-19.
Qian 2020 SARS-CoV-2 was detected in the rectum of a COVID-19 patient during the disease incubation period. There was direct evidence of replication of SARS-CoV-2 in the intestine.
Senapati 2020 This pilot study in India found SARS-CoV-2 RNA in fecal samples from 12 symptomatic and asymptomatic COVID-19 patients.
Tan 2020 In a single case report, SARs-CoV-2 was detected in the throat and rectum of the patient with COVID-19.
Tang 2020 An asymptomatic child was positive for a coronavirus by reverse transcription PCR in a stool specimen 17 days after the last virus exposure. The child was virus positive in stool specimens for at least an additional 9 days.
Wang Q-X 2020 Case series of 5 individuals who had Covid-19 and whose respiratory samples were negative by PCR, but had positive fecal samples. Observed over 3 to 15 days, no cases of reinfection occurred and all fecal samples became negative. Wu Y 2020 In 98 hospitalized Covid-19 cases, patients' faecal samples remained positive for SARS-CoV-2 for a mean of 11 days (maximum 5 weeks) after respiratory tract samples became negative.

Main findings of primary studies on orofecal transmission of SARS-CoV-2
Xiao F, Tang M 2020 39 of 73 hospitalized Covid-19 patients aged 10 months to 78 years tested positive for SARS-CoV-2 in fecal samples. Gastric, duodenal and rectal epithelial specimens collected via endoscopy of one patient were also studied: Immunofluorescence data showed that ACE2 protein, proven to be a cell receptor for SARS-CoV-2, is abundantly expressed in the glandular cells of gastric, duodenal, and rectal epithelia, supporting entry of SARS-CoV-2 into the host cells. Intracellular staining of viral nucleocapsid protein in gastric, duodenal, and rectal epithelia showed that SARS-CoV-2 infects these GI glandular epithelial cells.
Viral RNA was also detected in esophageal mucous tissue, but the absence of viral nucleocapsid protein staining in esophageal mucosa indicates low viral infection in esophageal mucosa. Detection of some viral nucleocapsid protein in rectal epithelial cells was observed in some additional Covid-19 patients, suggesting that some infectious viral particles may survive the GI environment.
Xiao F, Sun J 2020 This case series of 28 hospitalised patients for whom feces samples were available indicated that infectious virus was present in feces from two cases who also tested positive for viral RNA by RT-PCR.
Xing Y 2020 Three children showed a prolonged presence of SARS-CoV-2 in feces after throat swabs were negative.
Xu Y 2020 This study of 10 children with COVID-19 found that symptoms among children were nonspecific and relatively mild; rectal swabs tested positive among 8/10 cases even once nasopharyngeal tests became negative.
Yang 2020 Viral shedding and immunological features of 35 hospitalized children with Covid-19 were analyzed. 14/35 of the children had no symptoms; CT scan showed pneumonia in 32/35 children. Viral RNA was detected in fecal samples from 17/35. RNA was found in fecal samples up to 33 days after case detection.
Young 2020 SARS-CoV-2 Virus was detectable in the stool of 4 of 8 hospitalized patients.
Yuan 2020 A retrospective case note survey of 2,138 paediatric patients with suspected SARS-CoV-2 infection in Wuhan Children's Hospital included PCR tests on both throat swabs and anal swabs were available for 212 children. Viral loads detected on both throat and anal swabs available for 24 patients showed no significant difference. The findings suggested that in some children, fecal shedding may be a sign of prolonged mildly asymptomatic infection and represent the final phase of the disease.
Zhang J 2020 A small pilot sample of 14 hospitalised cases indicated agreement for the presence of COVID-19 between oropharyngeal samples and fecal samples.
Zhang T 2020 Three children with mild symptoms who were SARS-CoV-2 throat swab specimen negative on discharge from hospital were stool positive 10 days postdischarge.
Zhang W 2020 A small study of hospitalised COVID-19 patients indicated that RNA of SARS-CoV-2 may be shed via multiple bodily routes, and highlights that it is found in anal swabs sometimes when oral swabs show no viral RNA.
Zhang Y, Chen C, Zhu S 2020 A 2019-nCoV strain was isolated from a stool specimen of a laboratory-confirmed Covid-19 severe pneumonia case, who experienced onset on 16 January 2020 and was sampled on 1 February 2020 in China. The full-length genome sequence indicated that the virus had high-nucleotide similarity (99.98%) to that of the first isolated novel coronavirus isolated from Wuhan. In the Vero cells, viral particles with typical morphology of a coronavirus could be observed under the electron microscope.
Zhang Z, Chen C, Song Y 2020 Samples from 258 Covid-19 patients with clinical symptoms and positive PCR were collected: 93/258 stool samples were PCR positive; PCR-positivity in stool did not correlate with GI symptoms and only suggestively correlated with disease severity. Viral load tended to be higher within respiratory swab samples.
Live SARS-CoV-2 was isolated from a stool specimen (Ct value 24) of a severe Covid-19 patient (date of onset 16 January 2020) (strain HLJ002/HLJ/CHN/2020). The sequence of the full-length genome of strain HLJ002 indicated that the virus had high nucleotide similarity (99.98%) to the first SARS-CoV-2 (GenBank No. NC_045512) strain isolated from Wuhan. In the Vero cells, the virus caused obvious CPE, and the virus particles with typical morphology of coronavirus could be observed under the electron microscope. [NB it is not clear if additional cultures were attempted; we assume not.]

Sewage
Agrawal 2020 This study monitored the time course of the SARS-CoV-2 RNA concentration in raw sewage in the Frankfurt metropolitan area of Germany. 44 sewage samples were taken from three influent sources at two wastewater treatment plants, between April and August 2020. RT-qPCR was used to assess the presence and quantity of SARS-CoV-2 RNA. The correlation of this with concurrent epidemiological surveillance data was examined. Temporal dynamics were observed between different sampling points, indicating local dynamics in Covid-19 cases within the Frankfurt metropolitan area.
Ahmed 2020 Using samples collected between February and April 2020 from sewage treatment plants in Queensland, Australia, SARS-CoV-2 was detected by RT-qPCR assay, confirmed by sequencing.
Ampuero 2020 SARS-CoV-2 RNA was detected in untreated and treated wastewater samples obtained from two treatment plants in Santiago, Chile, March to June 2020.
Arora 2020 Untreated (influent), biologically treated, and disinfected wastewater samples were collected from May to August 2020 in two North Indian states; SARS-CoV-2 RNA was detected in 16/56 samples.
Betancourt 2020 In a US university campus, wastewater from a student dormitory was tested for SARS-CoV-2 RNA. Baseline tests established no SARS-CoV-2 when the students returned to campus; subsequently the virus' RNA was detected in wastewater samples for that dormitory. The students were isolated and tested by nasopharyngeal swab PCR to identify infected individuals. The study demonstrated surveillance using wastewater testing, leading to identifying and containing an outbreak.

Chavarria-Miro 2020
Testing of 24-hour composite raw sewage samples from two large wastewater treatment plants in Spain showed that SARS-CoV-2 was detected in sewage 41 days (15 January 2020) before the declaration of the first COVID-19 case (25 February 2020) in Spain, and in frozen samples dating back to 12 March 2019. If these results are confirmed, they suggest SARS CoV-2 has been circulating longer than first thought.
Curtis 2020 This study examined the variability of SARS-CoV-2 concentrations in wastewater grab samples collected every 2 hours for 72 hours compared with corresponding 24-hour flow-weighted composite samples. The results suggest that grab samples may be sufficient to characterize SARS-CoV-2 concentrations, but additional calculations using these data may be sensitive to grab sample variability and warrant the use of flow-weighted composite sampling.
Fongaro 2020 This study analysed human sewage in Florianopolis, Brazil from late October 2019 until the Brazil lockdown March 2020. SARS-CoV-2 was detected in two samples collected independently on 27th November 2019 (5.49 ± 0.02 log genome copies/L).

Fernandez-de-Mera
2020 This study investigated how readily SARS-CoV-2 RNA could be detected in environmental samples collected from an isolated small rural community in Spain at a time of a high COVID-19 prevalence (6% of the population of 883 inhabitants). Surface samples and village wastewater samples were taken: a number of these tested PCR-positive for SARS-CoV-2 RNA but two sewage samples tested negative.
Haramoto 2020 A study of the presence of SARS-CoV-2 RNA in wastewater and river water in a prefecture of Japan and compared two laboratory methods. Whilst 1 of 5 wastewater samples tested positive, no river samples tested positive for SARS-CoV-2 RNA.
Hata 2020 A study of wastewater samples over time in Japan reported that SARS-CoV-2 RNA detection frequency increased along with the number of reported cases, and was detected even at low prevalence of <1.0 per 100,000 people. Further, the detection frequency remained high even after the increase in cases stopped.
Iglesias 2020 This study measured SARS-CoV-2 RNA from a surface water source in a low-income settlement in Buenos Aires, Argentina between June and September 2020. Measurements of SARS-CoV-2 concentrations in surface water contaminated by sewage could be used to estimate changes in Covid-19 prevalence in the local community..

Lara 2020
This study in Belgium and the Netherlands investigated the use of phylogenetic analysis in routine wastewater testing samples to evaluate the diversity of SARS-CoV-2 at the community level, and compared these results with the virus diversity in patients. It showed that this method could approximate the diversity of SARS-CoV-2 viruses circulating in a community.
La Rosa 2020 b An environmental surveillance study based on twelve influent sewage samples, collected between February and April 2020 from wastewater treatment plants in Milan and Rome, Italy showed SARS-CoV-2 RNA fragments have been identified in sewage in Italy, and suggest a novel RT-PCR test for screening of waters.
Medema 2020 SARS-CoV-2 was detected in the sewage of five sites a week after the first COVID-19 case in the Netherlands. Even at low COVID-19 prevalence sewage surveillance could be a sensitive tool to monitor the viral circulation.

Main findings of primary studies on orofecal transmission of SARS-CoV-2
Neault 2020 In this longitudinal study, the stochastic variability inherent to wastewater-based epidemiology was corrected for using multiple fecal content protein biomarkers. These normalized SARS-CoV-2 protein data correlated well with public health SARS-CoV-2 prevalence metrics.
Ong 2020 The study ran from 24th January to 4th February 2020 and involved sampling in the physical areas around three COVID-19 patients at the Singapore dedicated SARS-CoV-2 outbreak center. The toilet bowl (seat and inner surface) and sink samples were positive, suggesting that viral shedding in stool could be a potential route of transmission. Post-cleaning samples were negative, suggesting that current decontamination measures were sufficient.

Peccia 2020
In an urban area of NE USA, this study of primary sewage sludge over time reported identifying SARS-CoV-2 RNA in all the samples. Adjusted for the time lag, the virus RNA concentrations tracked the Covid-19 epidemiological curve. SARS-CoV-2 RNA concentrations were a leading indicator of community infection ahead of compiled Covid-19 testing data and local hospital admissions.
Sharif 2020 78 wastewater samples collected from 38 districts across Pakistan, 74 wastewater samples from existing polio environmental surveillance sites, 3 from drains of Covid-19 infected areas and 1 from Covid-19 quarantine center drainage, were tested for presence of SARs-CoV-2. 21 wastewater samples (27%) from 13 districts were positive by RT-qPCR. This surveillance system has potential to aid monitoring of the pandemic, but attention is needed on virus concentration and detection assay to increase the sensitivity.

Shutler 2020
Combining in vitro data, pollution analysis and a virus survivability model, based on data from 39 countries, SARS-CoV-2 can remain stable within water for up to 25 days. Country-specific risk of infection posed by faecal contaminated water is environment-dependent, with water flow and temperature as important variables.
Trottier 2020 SARS-CoV-2 RNA was assessed in samples from the inflow point of the main waste water treatment plant of Montpellier, France, spring 2020. Samples were collected 4 days before the end of lockdown (7 May 2020) up to 70 days post-lockdown (20 July 2020). Increased amounts of SARS-CoV-2 RNA were observed from mid-June on, whereas the number of new Covid-19 cases recorded in the area started increasing a fortnight later.

Wang J 2020
The study reports the presence of SARS-Cov-2 in the hospital environment, surfaces, sewage, and the staff PPE in isolation wards in a Covid-19 hospital in China. SARS-Cov-2 RNA were positive from inlets of the sewage disinfection pool and negative from the outlet of the last sewage disinfection pool but no viable virus was detected by culture.

Wang XW 2020
No live SARS-CoV was found in any sewage samples from two hospitals receiving COVID-19 patients. SARS-CoV RNA was detected in sewage concentrates of two hospitals receiving SARS patients prior to disinfection, and occasionally after disinfection.
Wurtzer 2020 An increase of SARS-CoV-2 genome units in raw wastewaters in and around Paris, France accurately followed the increase of human COVID-19 cases observed at the regional level.
Zhao 2020 Wastewater, sludge, surface water, ground water, and soil samples of municipal and hospital wastewater systems and related environments in Wuhan during the Covid-19 middle and low risk periods were tested for SARS-CoV-2 RNA and the viral copies quantified using RT-qPCR. During the middle risk period, 1 influent sample and 3 secondary treatment effluents, 2 influent samples from wastewater system of a Covid-19 hospital were SARS-CoV-2 RNA positive. 1 sludge sample collected from a Covid-19 hospital 4 during a low risk period, was positive for SARS-CoV-2 RNA.

Toilet and or Sewage
Del Brutto 2020 SARS-CoV-2 prevalence and incidence were assessed in a rural Guatemalan village setting using serology. One month after baseline testing, 362 of 370 initially seronegative individuals were re-tested to assess incidence of seroconversion and associated risk factors. Twenty-eight of them (7.7%) became seropositive.
The overall incidence rate ratio was 7.4 per 100 person months of potential virus exposure (95%CI 4.7 to 10.2). The only covariate significantly associated with seroconversion was the use of an open latrine.

Ding Z 2020
This study randomly sampled in rooms and areas in the COVID-19 designated infectious diseases hospital Nanjing, China. 4/107 surface samples tested positive: two ward door door-handles, one bathroom toilet toilet-seat cover and one bathroom door door-handle. Three were weakly positive from a bathroom toilet seat, one bathroom washbasin tap lever and one bathroom ceiling exhaust louvre. 1/46 corridor air samples tested weakly positive.

Main findings of primary studies on orofecal transmission of SARS-CoV-2
Kang 2020 An outbreak of 9 confirmed cases of Covid-19 between 26 January 2020 and 13 February 2020 in 3 vertically aligned flats in a high-rise building in Guangzhou, China, during a period of social distancing, was investigated. There

Waterways
Guerrero-Latorre 2020 This study assessed the presence of SARS-COV-2 in urban streams from a low sanitation context i.e. highly impacted by sewage. Three river locations along the urban rivers of Quito, Ecuador were sampled on the 5 June 2020 during a peak of Covid-19 cases. SARS-CoV-2 RNA was detected in all samples, at levels similar to those in wastewater from cities during outbreaks. MERS-CoV has been shown to infect human primary intestinal epithelial cells, small intestine explants and intestinal organoids 8 . MERS-CoV has been detected in 42% of milk samples collected from lactating camels where it can survive for a prolonged period. A study of human primary intestinal epithelial cells and small intestine explants of MERS-CoV patterns identified the viral replication intermediates in stool specimens. MERS-CoV was found to be resistant to fed-state gastrointestinal fluids but less tolerant to the high acidic fasted-state gastric fluid.
Prolonged excretion of coronaviruses in feces was first observed in 1977 9 . In the SARS-CoV-1 outbreak in 2002-03, a significant portion of patients had enteric involvement. In the Toronto outbreak in 2003, 6% of 144 patients had diarrhoea on presentation 10 . Among 138 patients with SARS in Hong Kong, 20% presented with watery diarrhoea and 38% had symptoms of diarrhoea during the illness. Intestinal biopsy specimens showed the presence of active viral replication, and SARS-CoV RNA was detected in the stool of some patients for more than ten weeks after symptom onset 11 . A retrospective study on specimens from 154 patients in Hong Kong with laboratory-confirmed SARS found the viral load to be the highest in stool specimens 12 . Up to 70% of 75 patients in a community outbreak in Hong Kong developed watery diarrhoea 13 . This outbreak was linked to a faulty sewage system in the Amoy Gardens apartment complex, further suggesting orofecal transmission might be a route for transmission 14 .
The human gastrointestinal tract can act as a primary infection site for SARS-CoV. Ding et al. used a monoclonal antibody specific for the SARS-CoV nucleoprotein, and probes for the RNA polymerase gene fragment in four patients who died from SARS-CoV-1 15 . Virus was detected in the stomach, small intestine, distal convoluted renal tubule, sweat gland, parathyroid, pituitary, pancreas, adrenal, liver and cerebrum. The authors discussed that viruses in contaminated food and water may enter the human body through epithelial cells covering the surface of the gastrointestinal tract, although there was no direct evidence to show that food-borne transmission had occurred. A study from the sewage of two hospitals receiving SARS patients in Beijing found no infectious SARS-CoV contamination in any of the samples collected but did detect the nucleic acid in the sewage from the two hospitals before disinfection -providing further evidence that SARS-CoV-1 can be excreted by feces into the sewage system 16 .
Transmission of coronaviruses via the feces is established among animals: feline coronavirus, for instance, is typically shed in feces of healthy cats and transmitted by the orofecal route to other cats 17 . Pigs are also infected by the transmissible gastroenteritis coronavirus via the orofecal 18 . Bat coronavirus infects the gastrointestinal and respiratory tracts of bats, seemingly without causing disease 19 . Transmission following exposure to camel feces has also been considered biologically plausible 20 .
There is evidence that SARS-CoV-2 can survive adverse conditions in the gastrointestinal system. It has been identified in endoscopic specimens of the oesophagus, stomach, duodenum, and rectum of COVID-19 patients; substantial amounts of SARS-CoV-2 RNA have been consistently detected in stool specimens, and evidence suggests that SARS-CoV-2 can survive the adverse conditions in the gastrointestinal system. Heavy glycosylation of the large spike S protein has been shown to lead to resistance to the proteases, the low pH and bile salts found in the gastrointestinal system. Some gastric processes may actually facilitate viral entry into the enterocytes: in bovine coronavirus, one specific site on the S glycoprotein has to be cleaved by an intracellular protease or trypsin to activate viral infectivity and cell fusion 21 .
Evidence of ingestion, penetration of enterocytes and excretion of live SARS-CoV-2 is possible; however, this working hypothesis requires testing by conducting case-control studies during the investigation of outbreaks, following a set protocol. For such investigations, cases of COVID-19 (categorised by symptom presence and severity) either fecally excreting virions or not (cases and contacts) and controls would be healthy matches. Exposure to potentially fecally contaminated materials and protective measures taken would be elicited at interview. To minimise the play of recall and ascertainment bias, interviewers should be blind to fecal excretion status and the interview should take place as soon as possible after the event. Viability of fecal isolates and their possible pathogenicity should be tested in outbreaks.

Strengths and limitations.
This review is limited by the quality of included studies: many were small and did not provide a protocol that established a priori methods. Studies were often poorly reported and often did not take biases into account. Reporting is often heterogeneous and essential information such as symptom onset and cycle threshold values were often missing. We do not have information on publication bias, but the current urgency to understand SARS-CoV-2 may have an impact on research, with unknown implications and a tendency to publish those studies with positive results. We note the possibility that some cases may be reported in more than one publication, but there is no adequate method by which to identify this. It is likely we missed some studies, but we plan to keep updating this review. Furthermore, our judgments of quality are to some extended subjective and open to disagreement. This does not, however, undermine our overall assessment of the quality of the included studies. We perceive that standardization of methods in this area would improve the quality of the research. Some of these limitations increase uncertainties and prevent firm conclusions being drawn; however, this body of research provides largely consistent evidence on the main conclusions that SARS-CoV-2 is excreted fecally, is found in sewage and can be cultured from fecal samples.

Conclusion
Observational and mechanistic evidence as well as established animal orofecal transmission of coronaviruses suggests SARS-CoV-2 can infect and be shed from the human gastrointestinal tract. However, quantitative data on infectiousness and the consequent likelihood of transmission from orofecal contamination is not available. Whilst SARS-CoV-2 RNA is observed in sewage and wastewaters, there is no evidence of infectiousness in these sources and a transmission risk from these sources is considered unlikely based on the reviewed studies. To properly assess these risks, quantitative data on infectious virus are needed, along with information on likely infectious dose in humans.

Data availability
Underlying data All data underlying the results are available as part of the article and no additional source data are required. Furthermore, the authors and other readers of the manuscript may be interested to know that I recently published an invited review article in the Journal of Water and Health, a journal of the international Water Association, that is entitled: "Absence of virological and epidemiological evidence that SARS-CoV-2 poses COVID-19 risks from environmental fecal waste, wastewater and water exposures' (Sobsey, 2021 1 ).

Extended data
My JWH article further supports the assessments and conclusions of the authors of this F1000 Research article.
and extended data has been uploaded onto Figshare. This review is very well-written and the references relevant. The authors have undertaken this review thoroughly, with the largely lowmoderate data available. There are, however, some points that should be addressed to strengthen the review and conclusions.
The aim was to systematically review evidence on orofecal SARS-CoV-2 transmission while the conclusion focuses on SARS-CoV-2 fecal excretion and presence in sewage, which is not evidence of transmission. Detection of SARS-CoV-2 in stool samples and in ecological settings is important but it should be noted that was not the aim of this review. Viral culture data from six studies is included but this is a small minority of studies, and as the authors note due to methodological issues should be interpreted with caution.

○
Examining the ecological studies and review is very worthwhile but none address the issues of orofecal SARS-CoV-2 transmission. Perhaps ecological factors are a secondary outcome? ○ While the authors do note in the results that some reviews included overlapping studies, there is no figure/percentage given for this. How much overlap in total is there, is the overlap greater for cohort/case series or ecological studies? What is the quality of the overlapping studies -are they over-representing the low-quality data?
○ What is the breakdown for included reviews between cohort/case series and ecological studies? This is provided for primary studies in Figure 1 but I can't find it for the reviews.  Table 1 pages 10, 11 & 12 -total patient number from the cohort/case series (9081) appears in the "sewage" header rows.

Are the rationale for, and objectives of, the Systematic Review clearly stated? Yes
Are sufficient details of the methods and analysis provided to allow replication by others? Yes

Is the statistical analysis and its interpretation appropriate? Partly
Are the conclusions drawn adequately supported by the results presented in the review? Partly 3. Changed Discussion sentence: Many studies report identifying SARS-CoV-2 RNA in sewage and wastewaters, but viral culture from such sources has not been demonstrated, so there is no evidence of infection risk from those sources; however, detection of SARS-CoV-2 RNA in sewage and wastewaters can be useful as a surveillance tool.
4. Added: However, quantitative data on infectiousness and the consequent likelihood of transmission from orofecal contamination is not available. Whilst SARS-CoV-2 RNA is observed in sewage and wastewaters, there is no evidence of infectiousness in these sources and a transmission risk from these sources is therefore unlikely. To properly assess these risks, quantitative data on infectious virus are needed, along with information on likely infectious dose in humans.
5. Added to Discussion: We note the possibility that some cases may be reported in more than one publication, but there is no adequate means by which to identify this.
6. Unclear reporting in the reviews did not allow full analysis of their included studies. To do this we could revisit the full text of their included studies, but this is beyond the scope of this review. Inclusion of these reviews represents a level of interest in the field but necessarily they will overlap in scope and therefore we do not analyse their findings in detail.
7. Thank you for the suggestion to further sub analyse the data; at this stage there are insufficient data for such sub analyses including among children.
8. Unfortunately there are several studies in which sex of participants is not reported and so we were limited by this.
9. This was a very interesting suggestion to sub-analyse according to preprint publication and examine possible publication biases; this is unfortunately beyond the scope of this review, not least because this changes over time. The next time we update this review we hope to have a better breakdown of those fully published, and hopefully a better bank of from human exposure. This point should be mentioned in the overall presentation of the infectious virus results. Furthermore, we do not know the infectivity dose-response relationship of SARS-CoV-2 for human hosts, such an estimate of 50% infectious dose. Unless we know what virus concentration is in samples such as feces or sewage/wastewater and unless we know what the likelihood of human infection is from exposure to known concentrations of infectious virus, it is probably unjustified to suggest that the presence of an unknown concentration of infectious virus in a very small number of reported virus-positive fecal samples poses a human health risk from exposure. While infectious virus presence in a few samples is noteworthy, the lack of quantitative infectivity data makes it impossible to judge that there is plausible human health exposure risk from this infectious virus-positive fecal matter of unknown concentration. The best one can say is that the risk of infection or other health effects from exposure is not zero, but is also not quantifiable due to lack of data on virus concentration in feces and lack of knowledge of human infectivity dose-response for infection or other health endpoints.
For sewage and water, the lack of data on virus infectivity in such samples makes it impossible to determine if these samples pose risks of human infection from exposure. Although some data on concentrations of viral RNA are reported based on estimated gene copies or CT values, the quantitative relationship of virus infectious units to RNA concentration units is highly variable and therefore uncertain This is especially so for samples that have been in the environment for unknown and variable periods of time since the feces were shed from a human host. This is due to the variable and difficult-to-quantify effects of exposure time, temperature, microbial activity, predation and other environmental stressors on virus survival. These factors are not mentioned in the paper, even though we know they are important for determining the quality and validity of the results for infectious virus presence and concentrations. The last sentence of the first paragraph of the Discussion section does not adequately address the issue of lack of evidence of virus culture in wastewater and sewage and its implications for human health risks from sewage. Instead it sidesteps to the potential value using SARS-CoV-2 RNA in sewage and wastewater as a surveillance tool. This seems to miss the main point of assessing the evidence for human health risk from exposure to sewage or wastewater.In the "strengths and limitations section of the Discussion section, the authors state "however, this body of research provides largely consistent evidence on the main conclusions that SARS-CoV-2 is excreted fecally, is found in sewage and can be cultured from fecal samples." While there there is some evidence that unknown but probably low concentrations of infectious SARS-CoV-2 have been occasionally found in excreted feces, they have not been found in sewage or wastewater. The wording of the above sentence conflates feces with wastewater and sewage, even though there is only limited and poor evidence for infectious virus in the latter and no evidence for it in sewage or wastewater. I recommend that the sentence be revised to avoid conflating these rather separate exposure sources of feces versus sewage and wastewater.

3.
The Conclusion sentence is short and limited in scope. It does not adequately address the issue of whether sufficient infectious SARS-CoV-2 is or is not likely to be present in feces or wastewater and sewage to pose human health risks from exposure. The available evidence seems to indicate that such risks are not possible to assess or quantify without better and more rigorous and as well as more quantitative data that is still not available.

4.
these risks, quantitative data on infectious virus are needed, along with information on likely infectious dose in humans.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias • You can publish traditional articles, null/negative results, case reports, data notes and more • The peer review process is transparent and collaborative • Your article is indexed in PubMed after passing peer review • Dedicated customer support at every stage • For pre-submission enquiries, contact research@f1000.com