As no curative treatment for SARS-CoV-2 is currently available, most public health measures to contain the pandemic are based on preventing the spread of the pathogen. The virus is transmitted by the respiratory route and by direct contact with contaminated surfaces and subsequent contact with nasal, oral or ocular mucosa. Although patients with symptomatic coronavirus disease 2019 (COVID-19) have been the main source of transmission, observations suggest that asymptomatic and incubating patients also have the ability to transmit SARS-CoV-2. Angiotensin II converting enzyme (ACE2) is the main cellular receptor for SARS-CoV-2, which interacts with the spike protein to facilitate its entry. ACE2 receptors are highly expressed in the oral cavity and present at high levels in oral epithelial cells. The mean expression of ACE2 was higher in the tongue compared to that in other oral tissues and it has been found to be higher in the minor salivary glands than in the lungs. These findings strongly suggest that the oral cavity and specifically the saliva may be a high-risk route for SARS-CoV-2 infection. Thus, strategies reducing salivary viral load could contribute to reduce the risk of transmission. Furthermore, studies using macaques as animal models have shown that SARS-CoV persists for two days in oral mucous membranes before its diffusion to the lower respiratory tract. This offers an interesting preventive and therapeutic window of opportunity for the control of this disease. For this reason, the use of mouthwashes with antiseptics that have virucidal activity can be a simple preventive strategy that could easily be applied both by infected patients before being examined by sanitary personnel and in the general population. This study is a multi-centered, blinded, parallel-group, placebo-controlled randomised clinical trial that tests the effect of three different mouthwashes (chlorhexidine 0,12%, chlorhexidine 0,2%, and Cymenol -ZnCl2) in the salivary viral load of SARS-CoV-2 measured by qPCR at three different timepoints. A fourth group of patients using a distilled water mouth rinse is used as a control. Viral particles per ml of saliva are quantified at baseline and 5, 15 and 60 minutes after a 1-minute mouth rinse with the antiseptic or water. The study aims to test whether any of these standard oral antiseptics appear to diminish viral load in saliva and could therefore be used as a strategy to reduce transmission risk in clinical and social settings.

Strategies reducing salivary viral load could contribute to reduce the risk of transmission of SARS-CoV-2. Thus, the use of mouthwashes with antiseptics that have virucidal activity can be a simple and low-cost preventive strategy that could easily be applied in the general population. To evaluate the effect of several antiseptics to neutralize or reduce the SARS-CoV-2 viral load in vivo in saliva samples, a multicentre, randomized, blind, four-parallel-group, placebo-controlled trial has been designed.

The study will be performed in Madrid, in three different hospitals: Fundación Jiménez Díaz University Hospital (Madrid, Spain), Villalba University General Hospital (Madrid, Spain) and Infanta Elena University Hospital (Madrid, Spain).

Every patient included is previously diagnosed and hospitalized because of SARS-COV-2 infection, being admitted mainly for respiratory pathology. All of them will be adults (age >18 years) and provide their voluntary written or oral consent to participation according to the hospital's ethics committee recommendations.

After approval of consent, the hospital staff responsible for the interventions will consecutively assign each participant a code following the order from a previously randomly generated table. The code will consist of a patient number and a letter corresponding to one of the four study groups (A, B, C and D), that will be known to the clinical personnel but unknown to the laboratory personnel who will process samples and extract RNA, as well as to those that will analyse the data. In this way, participants will be randomly assigned to one of the four treatment groups and the blind will be achieved by using identical tubes with the same volume for both mouthwashes and placebo.

Every included patient is asked not to eat, drink anything but water, chew gum, smoke, brush their teeth, or use any mouthwash for one hour prior to sample collection. In addition, they are not allowed to drink for half an hour after the mouthwash and eat for the entire test.

Four mouthwashes are randomized: rinses are ready to use in their commercial formulas. All mouthwashes are commercial products classified as safe.

A total of 4 non stimulated saliva samples will be collected for each patient: one basal and three after the mouthwash, at 5 minutes, at 15 minutes and at 60 minutes, respectively. Each participant was asked to donate four saliva samples of at least 2 mL: one at baseline before rinsing, and other three at 5, 15 and 60 minutes after mouthwash. Each sample provided by the patient was collected using the drooling technique, discarding secretions obtained after expectoration in a sterile millimetre plastic tube. After collecting the baseline saliva sample (pre-rinse, t0), each patient rinsed with the randomised mouthwash (15 mL) for one minute. The patient was instructed to ensure that the mouthwash reached all teeth, gums, palate and tongue, without swallowing or gargling. At 5 minutes (t5'), 15 minutes (t15') and 60 minutes (t60') after rinsing, respectively, at least 2 mL of saliva were collected per sample. Once the required 2 mL of saliva sample had been reached, the plastic tube was labelled with the code assigned to the patient and the time point, and kept hermetically sealed at -80ºC. All samples were transferred to the Biosafety Level 3 (BSL-3) laboratory (FISABIO-Public Health) on dry ice according to UN3733 standards.

In the BSL-3 laboratory at FISABIO-Public Health, the samples were thawed at room temperature and a 200 uL aliquot of each was used for RNA extraction. The remaining volume of each sample was immediately stored at -80C until further testing in Vero-E6 cell lines. For RNA extraction, firstly, lysis was performed with proteinase K (Epicentre) for 20 minutes at 56°C followed by the fully automated eMAG platform (bioMérieux, France) according to the manufacturer's instructions for saliva samples. Then, to detect the SARS-CoV-2 E gene, a multiplex RT-qPCR test was performed based on WHO-Charité and U.S. CDC assays(13,14), together with the human RNAse-P gene as a sample and extraction control following the details of the protocol described by Ferrer et al.(15). For the multiplex RT-qPCR two replicates per sample of extracted RNA were performed. Results were normalised to virus copies per mL of saliva.

For SARS-CoV-2 culture from saliva samples in Vero-E6 cells, Vero-E6 cells (ATCC) were cultured in supplemented DMEM (Biowest) following the details of the protocol described by Sánchez Barrueco et al.(15).

Saliva samples diluted 1:1 in 1X Dulbecco's PBS (Gibco) were centrifuged (5 minutes at 12,000 g) and 300 uL of the supernatant was incubated in duplicate with 1.5x105 Vero-E6 cells in a 24-well plate (Corning) for one hour at 37°C to allow absorption of live viruses. Then, saliva (with unabsorbed viruses) was removed and replaced by 500 uL of infection media (supplemented DMEM with trypsin TPCK 6ug/ml (Biowest)) and incubated for 5 days at 37°C and 5% CO2. After 5 days of infection, both supernatants were collected for RNA extraction following the eMAG platform instructions and cytopathic effect (CPE) was recorded as negative or positive(14). CPE was evident in positive controls-cells seeded with 300 uL SARS-CoV-2 stock virus (MAD-6 strain, 1.33 TCID50/mL, CNB-CSIC, Spain)-, absent in negative controls-cells seeded with 300 uL PBS- and variable in the cells seeded with saliva samples, depending on the viral load in saliva. A culture was considered positive when the RT-qPCR Ct value in day 5 culture supernatant was < 37 (equivalent to ≥ 2x103 SARS-CoV-2 copies/mL).

Regarding the analyses to be performed, the main objective of the study is to determine the effect of the different mouthwashes on the SARS-Cov2 viral load in saliva tested in vivo. Thus, the primary outcome will be the change in salivary viral load viability between baseline and the three post-mouthwash timepoints in each treatment. In addition, the correlations between the basal viral load and viability prior to rinsing with the different clinical data collected, such as age, sex, days since the appearance of symptoms and days since the patients' virus positivity determined by PCR of nasopharyngeal samples will also be studied. Finally, the possible associations between categorical clinical variables and the frequency of patients who improve in viral load and viability at different times for the different treatments will be evaluated.

Assuming that the baseline saliva sample is considered as a control for each patient, the investigators considered reaching 5 patients per arm (CHX 0.12%, CHX 0.2%, Cym ZnCl2 and placebo) to identify significant differences between groups of more than 20%. With a Cronbach's alpha of 0.05, power of 0.8 and a 10% possible drop-out rate due to drop-out or insufficient viral load, a minimum sample size of 22 patients was calculated.

The Wilcoxon test was used to assess mean differences between study time points (baseline and after 5, 15 and 60 minutes) and between different mouthwashes or study arms. Spearman's correlation analyses was performed to assess relationship between clinical variables with baseline salivary viral load, viable load, or nasopharyngeal viral load. All calculations and tests were performed with R software (version 3.6.3, "stats" package)

Inclusion Criteria:

• Lower than 7 days from the positive SARS-COV-2 PCR test of a nasopharyngeal sample
• Have the ability to donate saliva samples and perform mouthwash

Exclusion Criteria:

• Patient participation in a COVID-19 research study using experimental drugs,
• Use of an antiseptic mouthwash for 48 h before the start of the study
• Any known hypersensitivity or allergy to components of the mouthwashes.

The hospital staff consecutively assigned each participant a code following the order from a previously randomly generated table.

The code were unknown to the personnel who processed the samples and to those that analyzed the data.

Participants were blinded by using identical tubes with the same volume for both mouthwashes and placebo.