Faecal Viral Excretion and Gastrointestinal Co-Infection Do Not Explain Digestive Presentation in COVID-19 Patients

The physiopathological mechanisms responsible for digestive symptoms in COVID-19 patients are still unclear. The aim of this study was to determine the influence of faecal viral shedding on digestive symptoms and propose differential diagnoses in order to understand the gastrointestinal clinical spectrum in acute cases of COVID-19. All patients managed between March and May 2020, from whom stool samples were collected for microbiological investigations, were included. Microbiological analysis consisted of syndromic PCR screening and microscopic parasitological examination supplemented with microsporidia and multiplex protozoa PCR. SARS-CoV-2 infection was diagnosed via viral detection in respiratory and frozen stool samples, completed via serological test when necessary. Epidemiological, clinical, radiological, and biological data and clinical courses were compared according to COVID-19 status and faecal SARS-CoV-2 shedding and enteric co-infection status. The sample included 50 COVID+ and 67 COVID− patients. Faecal viral shedding was detected in 50% of stool samples and was associated with a higher viral load in the upper respiratory tract. Detected enteric pathogens were not different between subjects with different COVID-19 statuses or faecal SARS-CoV-2 shedding and had no impact on the clinical course for COVID-19 patients. The connection between SARS-CoV-2 shedding and enteric pathogen co-infection involvement in gastrointestinal presentation and clinical course is still unclear, suggesting other processes are involved in digestive disorders in COVID-19 patients.

While the gastrointestinal tropism of SARS-CoV-1 and MERS-CoV is well-described, the mechanisms underlying digestive disorders in SARS-CoV-2 are still under investigation. Similar to SARS-CoV-1, SARS-CoV-2 penetrates alveolar, digestive and hepatic cells via interactions between the viral spike S glycoprotein and the host receptor angiotensinconverting enzyme 2 (ACE2), followed by membrane fusion enhanced by the co-expressed serine proteases TMPRSS2 [12]. Previous studies reported the presence of intracytoplasmic nucleocapsid in digestive biopsies and prolonged viable viral particle shedding in stool samples from COVID-19 patients, suggesting intestinal viral multiplication and a potential risk of faecal-oral transmission of SARS-CoV-2 [13][14][15][16]. However, the relationship between SARS-CoV-2 faecal shedding and digestive disorders is still unclear [9,17]. In addition to the hypothesis of a direct effect of the virus on gastrointestinal presentation, the influence of host-related or infection-related factors has also been suspected [18]. Alterations of the gut microbiome via metabolic and/or inflammatory mechanisms have also been observed in SARS-CoV-2 patients [19,20], and the resultant dysbiosis was thought to facilitate and sustain gastrointestinal involvement via opportunistic pathogen proliferation, intestinal inflammation, and metabolic disorders [18][19][20][21][22]. Although early observations of COVID-19 patient cohorts in China and later in the USA and Europe considered SARS-CoV-2 the main causative agent in these intestinal disorders, few studies included a comprehensive investigation of the differential diagnosis of gastrointestinal disorders described in COVID-19 patients [23,24].
During the first French epidemic wave in the spring of 2020, SARS-CoV-2 faecal shedding, as well as potential gastrointestinal co-infections during the clinical course of COVID-19 patients, were investigated in a French hospital in order to better identify the role of SARS-CoV-2 infection in digestive symptoms and to optimise clinical management.

Ethical Statement
Ethical review and approval were waived for this study due to thatthis single centre study was performed according to current French legislation for the protection of individuals during health care management (https://www.legifrance.gouv.fr/codes/article_lc/ LEGIARTI000045629992/ (accessed on 22 April 2022)) and used existing medical data. No informed consent was required for the study other than the lack of refusal by the patient to participate in medical research. Information about the possibility of using patients' medical data, unless they express their opposition, is present in their medical reports. In cases where patients refused permission to use their data, those patients were not included.

Study Population
This prospective study was conducted from 1 March 2020 to 3 May 2020 during the first wave of the SARS-CoV-2 pandemic in France. All patients receiving treatment at Bichat-Claude Bernard Hospital who provided stool samples for microbiological investigations were included, regardless of the reason for care. For patients with suspected COVID-19, SARS-CoV-2 molecular diagnostic tests were performed on respiratory samples on admission, with additional SARS-CoV-2 molecular detection in stool samples. In cases where the clinical picture was strongly suggestive of COVID-19 (typical clinical presentation and history and radiographic detection of ground-glass opacities) but respiratory samples were negative, SARS-CoV-2 serology in the month following the acute infectious episode was performed to confirm COVID-19 (COVID+ group). Patients with SARS-CoV-2 RNA detected in their stool samples were assigned to the excretory group as opposed to patients with undetectable faecal viral shedding (non-excretory group). Patients who presented an atypical clinical presentation and history without typical pulmonary radiography of COVID-19 and/or negative results for SARS-CoV-2 PCR on respiratory samples were retrospectively and prospectively included, according to the above-mentioned criteria, from 1 February 2020 to 3 May 2020, to constitute a negative control group for gastrointestinal co-infections (COVID− group). Nosocomial COVID-19 infection (i.e., proven SARS-CoV-2 infection acquired during hospital stay) and onset of digestive symptoms later than 30 days after hospital care were excluded. For each included patient, epidemiological (age, sex, place of residence, travel history), current treatment (background and anti-infectious treatment), clinical (underlying disease, disease history, type and duration of clinical signs, clinical course and outcomes), radiological (extent of pulmonary ground-glass opacities), and biological data (sampling, microbiological, haematological and biochemical investigations) were collected from medical records.
The aetiology of digestive symptoms was determined by assessing current medical treatment, underlying disease, disease history, type and duration of gastrointestinal symptoms and microbiological results. When possible, the onset and duration of digestive and respiratory symptoms, as well as the interval between the onset of symptoms and the start of medical care, the delay between digestive and respiratory symptoms, length of hospital stay, respiratory and stool sampling (i.e., the time between the onset of clinical symptoms and biological sampling) as well as the delay between digestive and respiratory sampling were extracted from patient records. ICU management was taken into account from at least day 1 of hospitalisation in the unit. The collected underlying diseases were chosen because they had already been considered as risk factors for severe COVID-19 [3,4]. COVID-19 was considered severe if infected patients were managed in the ICU and/or if they died. Fever was defined as a body temperature over 38 • C. Diarrhoea was defined as a modification of stool consistency associated with an increased frequency of bowel movements. Diarrhoea was considered chronic if it persisted for at least 30 days.

Virological Investigations
The RealStar ® SARS-CoV-2 RT-PCR Kit 1.0 assay (Hamburg, Germany), targeting the SARS-CoV-2 N and E genes was used for viral detection. RNA extraction was performed with the MagNA Pure LC 2.0 System (Roche, Basel, Switzerland) according to the manufacturer's protocol. For each sample, 200 µL of nasopharyngeal (NP) sample or faeces suspension was diluted in 2 mL of lysis buffer and eluted in 50 µL. A total of 10 µL of extracted RNA was used to perform the real-time RT-PCR with an ABI Prism ® 7500 SDS (Applied Biosystems, Waltham, MA, USA). A heterologous amplification system (Internal Control), included in the kit, was used to identify possible RT-PCR inhibition. In case of inhibition, NP and faeces samples were retested under the same conditions at a 1:10 dilution in PBS. SARS-CoV-2 anti-N antibody detection was performed using Abbott Architect SARS-CoV-2 immunoglobulin G (IgG) (Abbott, Maidenhead, UK) and expressed as an index (cutoff: 0.49).
Trained microscopists also performed microscopic parasitological examinations on the same stool sample after iodine-stained wet mount, flotation, and biphasic concentration. In addition, stool smears stained by modified Ziehl-Neelsen, Microsporidia PCR (Bioevolution ® , Bussy-Saint-Martin, France) and a 4-plex protozoa PCR targeting E. histolytica, Cryptosporidium spp., G. intestinalis and D. fragilis (Amplidiag ® , Mobidiag ® , Espoo, Finland) were also systematically performed according to the manufacturer's instructions. DNA was extracted from stool on EZ1 ® Advanced XL (Qiagen ® ) using the EZ1 ® DNA tissue kit and the EZ1 bacteria card with PBS suspensions of stool samples pre-treated with proteinase K.
A diagnosis of gastrointestinal co-infection was recorded if an enteric microorganism, known as a possible pathogen, was detected in a patient who presented digestive symptoms without other evident aetiology combined with a high microorganism burden (positive microscopic examination and/or Ct < 30 on QIAstat-Dx ® and/or Ct < 35 on Amplidiag ® ). Otherwise, carriage of the detected microorganisms was considered.

Statistical Analysis
Quantitative variables were represented by the mean and standard deviation or by the median and interquartile range [25th percentile-75th percentile] according to their distribution. Categorical variables were evaluated according to size and frequency. Fisher's exact test was used to compare categorical variables between groups. Kruskall-Wallis tests were used to compare quantitative variables between groups when appropriate. For all tests, a difference was considered significant when p < 0.05. All reported p values are two-tailed. Statistical analyses were performed using STATA, version 12 (Stata corp ® , College station, TX, USA).

Differential Diagnosis of Gastrointestinal Disorders
Overall, after reviewing the individual medical charts of the included patients, underlying diseases could be the cause of gastrointestinal symptoms in 30% of patients (20/67) in the COVID-group and 4% (2/50) in the COVID+ group (p < 0.01) whereas iatrogenic aetiology was suspected in 20.9% (14/67) in the COVID-group but not in COVID+ patients (p < 0.01). Aetiology remained unclear in only 6% of COVID-patients (4/67).

Differential Diagnosis of Gastrointestinal Disorders
Overall, after reviewing the individual medical charts of the included patients, underlying diseases could be the cause of gastrointestinal symptoms in 30% of patients (20/67) in the COVID− group and 4% (2/50) in the COVID+ group (p < 0.01) whereas iatrogenic aetiology was suspected in 20.9% (14/67) in the COVID− group but not in COVID+ patients (p < 0.01). Aetiology remained unclear in only 6% of COVID− patients (4/67).
Presumed food contamination was suspected in 32% and 37% of COVID+ and COVID− groups (p = 0.7) whereas faecal-oral transmission of gastrointestinal microorganisms was suspected in 100% and 93% respectively (p = 0.5). The number of stools with enteric pathogens, the total number of detected enteric pathogens, the number of pathogenic, commensal or total microorganisms (bacteria, virus and parasite) detected and the number of each pathogenic, commensal or total detected species were compared between COVID+ and COVID− patient groups (Figure 2) as well as between excretory and non-excretory RNA SARS-CoV-2 patients and no difference was observed in these patient group for the tested variables. Moreover, no influence of the detected gastrointestinal microorganisms was observed on either clinical course or outcomes in COVID-19 patients (Table 3). Presumed food contamination was suspected in 32% and 37% of COVID+ and COVID-groups (p = 0.7) whereas faecal-oral transmission of gastrointestinal microorganisms was suspected in 100% and 93% respectively (p = 0.5). The number of stools with enteric pathogens, the total number of detected enteric pathogens, the number of pathogenic, commensal or total microorganisms (bacteria, virus and parasite) detected and the number of each pathogenic, commensal or total detected species were compared between COVID+ and COVID-patient groups (Figure 2) as well as between excretory and nonexcretory RNA SARS-CoV-2 patients and no difference was observed in these patient group for the tested variables. Moreover, no influence of the detected gastrointestinal microorganisms was observed on either clinical course or outcomes in COVID-19 patients (Table 3).

Discussion
SARS-CoV-2 is an emerging respiratory virus that preferentially targets the upper respiratory tract and spreads extensively through aerosol transmission and direct contact [25]. Several observations have also implied SARS-CoV-2 in digestive symptoms potentially explained by gastrointestinal tract tropism as already described for SARS-CoV-1 [12]. However, the inconsistent presence of gastrointestinal symptoms in COVID-19 and the lack of knowledge of the underlying physiopathological mechanisms of gastrointestinal infection by SARS-CoV-2 challenge the idea of direct viral involvement in digestive complications. Differential diagnosis of gastrointestinal symptoms in COVID-19 patients is warranted to better understand the spectrum of this viral infection and to optimise clinical management with targeted therapy [23,24]. The aim of the work described here was to explore the involvement of viral shedding in gastrointestinal symptoms and possible differential diagnoses in COVID-19 patients managed in our hospital centre during the first wave of the pandemic in France.
Despite specific hospital organisation during the first wave of SARS-CoV-2, with hospital care management only for symptomatic patients at-risk of severe COVID-19, our patient population is consistent with previous Asian and US patient cohorts [9]. Gastrointestinal presentation was observed in patients at risk of developing severe COVID-19 (i.e., with higher BMI, high incidence of hypertension, elevated SGPT levels and low lymphocyte count) with a significant mortality of 18% [7][8][9][10][11]26]. The early onset and short duration of digestive signs observed in the COVID+ population during clinical management, associated with systemic symptoms such as fever, myalgia and ageusia/anosmia, is in line with COVID-19 acute digestive episodes. In addition, only COVID− patients with severe clinical conditions were hospitalized during the first wave of the SARS-CoV-2 pandemic, which could explain the more prolonged digestive presentation more frequently resulting from underlying diseases or associated medications.
Faecal viral shedding has no impact on digestive presentation and clinical course in COVID-19 patients as previously observed in Chinese and US cohorts [6,8,9,17,27] challenging the hypothesis of a direct viral cytolytic effect on infected enterocytes. However, faecal SARS-CoV-2 excretion seems to represent a severity marker as stools positive for SARS-CoV-2 RNA were more frequently observed in patients with higher numbers of underlying diseases associated with severe COVID-19, and correlated with a high viral load in upper respiratory samples resulting from lack of control of viral multiplication by innate immunity in the population at risk of severe COVID-19 [18,27,28]. Although a small number of studies have demonstrated the presence of viable viral particles in stool and in enterocytes of COVID-19 patients [13][14][15][16], our results could not exclude passive passage of altered SARS-CoV-2 through the carriage of respiratory mucus from the upper respiratory tract to the gut [15]. The most frequent detection of SARS-CoV-2 in patients reported sputum production is also an argument for protected passive viral passage through swallowing mucus [15,16]. Therefore, faecal viral shedding may more likely reflect a high respiratory viral load as opposed to viral multiplication in the gastrointestinal tract. An indirect role of viral particles in stool was also suspected by inducing intestinal pro-inflammatory IL-18 response leading to altered intestinal microbiota, systemic inflammation, and the onset of a cytokine storm [29]. Our study's focus was on digestive symptoms compared between small excretory and non-excretory groups which could explain the absence of differences between the variables assessed. Additional data would be necessary to explore these hypotheses further.
Gastrointestinal infectious diseases seem to be the main cause of the observed digestive symptoms in COVID-19 patients. Differential microbiological diagnostic revealed that, among the 38% of SARS-CoV-2 stool samples containing enteric microorganisms, approximately 20% were co-infected with known gastrointestinal pathogens. The incidence of gastrointestinal pathogens was similar between COVID+ and COVID− patients, in contrast with previous studies which reported a higher incidence of C. difficile in COVID+ patients [24]. One of the physiopathological hypotheses for enteric symptoms during COVID-19 is an alteration of the intestinal microbiome through dysregulation of the lung-gut axis promoting a dysbiosis, characterised by a reduction in the diversity, an initiation of an inadequate systemic and local immune response and an alteration in the host's metabolism, which contribute to the maintenance of digestive damage and proliferation of opportunistic intestinal pathogens [18][19][20][21][29][30][31][32]. Similar distributions of enteric commensal and known pathogenic bacteria, viruses, and parasites between COVID and excretory groups were observed, indicating, firstly, that COVID-19 patients are not overexposed to faecal-oral transmission compared to our control group and, secondly, that, SARS-CoV-2 infection and faecal shedding do not specifically promote gastrointestinal pathogen infection. However, we noticed trends concerning the larger panel of gastrointestinal pathogens detected in COVID− patients compared to COVID+ patients (18 different pathogens versus 10 in COVID+ patients) (Figure 2). Loss of protozoan and helminth pathogens coupled with a non-significant increase of Blastocystis spp., D. fragilis, EAEC, EPEC and C. difficile from COVID− to COVID+ patients could be explained by the initiation of an altered gut microbiome associated with intestinal inflammation. The presence of non-pathogenic amoeba (Blastocystis spp., D. fragilis, Entamoeba spp., Endolimax nana), helminths, or Cryptosporidium spp. in the gut was initially considered beneficial for the host, contributing to the diversity of bacteria microbiota and to an intestinal anti-inflammatory response [33][34][35]. By contrast, EPEC, EAEC, as well as the intestinal proliferation of Blastocystis spp. and D. fragilis, have been associated with the irritable bowel syndrome (IBS) [36]. Complementary studies would be necessary to confirm a possible IBS-like gastrointestinal presentation during severe COVID-19. Microbiological investigations for differential diagnosis are warranted in cases of gastrointestinal presentation during SARS-CoV-2 infection to propose targeted antiinfectious therapy against frequent community-acquired diarrhoea. However, in our study context, no impact of gastrointestinal microorganisms was observed on the clinical course of COVID-19 patients and on SARS-CoV-2 shedding in stool. The first possible explanation is methodological, as the small size of the study group and the low number of enteric pathogens detected, resulting from incomplete microbiological investigations, did not allow us to identify differences between the groups. The second possible explanation is pathophysiological, with a possible predominant effect of dysbiosis induced by COVID-19 compared with gastrointestinal infection by conventional pathogens. This latter hypothesis would corroborate our previous observations on the lack of impact of faecal viral excretion on the gastrointestinal presentation of COVID-19.
This work was designed as a preliminary prospective single centre proof-of-concept study during the first wave of the COVID-19 pandemic in France to explore the frequency and impact of differential gastrointestinal diagnoses in COVID-19. This design presented some limits. The particular context in which this work was performed during the first wave of the COVID-19 pandemic in France precluded the inclusion of a large numbers of patients, and exhaustive microbiological investigations were not possible. Owing to its single-centre design, microbiological investigations could have biased the panel of enteric pathogens identified. The microbiological approach was based in most cases only on one stool sample, possibly underestimating the diagnosis of parasites. These different factors could have contributed to the lack of differences between patient groups. Moreover, we were not able to recruit non-severe COVID+ and COVID− patients with different underlying diseases which also limited the possibilities of observing the variety in number and types of gastrointestinal pathogens and their impact on clinical course, intestinal microbiome and outcome between patient groups. In order to assess the hypothesis of a systemic impact of dysbiosis during COVID-19, other clinico-biological variables should be chosen to compare patient groups, such as, for example, the lack of data collected regarding co-morbidity factors, the intensity of the digestive presentation or detailed care management, which could also help to explain the absence of difference between groups, emphasising the need for further investigation. Clinical management and anti-infectious therapy were not standardised in this study, leaving it to the discretion of the clinicians, thereby adding heterogeneity to data on clinical course and outcome.
This study highlights the continued ambiguity concerning the nature of viral physiopathological mechanisms in COVID-19 digestive presentation. Whereas faecal RNA viral shedding could simply reflect the passive transfer of high nasopharyngeal viral loads, an indirect viral effect through alteration of gut microbiome seems to be a more plausible explanation considering both the absence of any observable impact of SARS-CoV-2 faecal shedding and gastrointestinal co-infection on the clinical course, and the nature of gastrointestinal pathogens detected in COVID-19 patients associated with an IBS-like presentation. Therefore, the importance of differential microbiological diagnosis in investigating any digestive disorders present in COVID-19 patients appears warranted, since the levels of detection of classic enteric pathogens were similar in non-COVID-19 patients with enteric symptoms. Data Availability Statement: Authors confirm to the editorial microorganisms board that all materials and raw data necessary for manuscript reviewing and comprehension will be from the corresponding author when necessary.