SARS-CoV-2 RNA in exhaled air of hospitalized COVID-19 patients

Knowledge about contagiousness is key to accurate management of hospitalized COVID-19 patients. Epidemiological studies suggest that in addition to transmission through droplets, aerogenic SARS-CoV-2 transmission contributes to the spread of infection. However, the presence of virus in exhaled air has not yet been sufficiently demonstrated. In pandemic situations low tech disposable and user-friendly bedside devices are required, while commercially available samplers are unsuitable for application in patients with respiratory distress. We included 49 hospitalized COVID-19 patients and used a disposable modular breath sampler to measure SARS-CoV-2 RNA load in exhaled air samples and compared these to SARS-CoV-2 RNA load of combined nasopharyngeal throat swabs and saliva. Exhaled air sampling using the modular breath sampler has proven feasible in a clinical COVID-19 setting and demonstrated viral detection in 25% of the patients.

Participants. Patients ≥ 18 years of age were included prospectively within 7 days after admission to the Radboudumc. COVID-19 was confirmed by RT-qPCR on a combined nasopharyngeal throat swab.
Sample collection. Patient characteristics, clinical features, and routine hemocytometric and inflammatory laboratory measurements were collected on standardized patient charts. Routine diagnostic laboratory measurements were registered on the day of sampling (+ /− 1 day). From each patient, samples were simultaneously collected using three different methodologies. A combined nasopharyngeal throat swab was taken using the hospital's protocol according to national guidelines 12 . After collection, the sample was placed in 5.0 mL virus transport medium consisting of Hank's balanced salt solution (Gibco) containing 2% FCS (Sigma-Aldrich), 100 µg/mL gentamicin (Gibco), and 0.5 µg/mL amphotericin-B (Gibco) and stored at -20 °C until further processing.
Exhaled air was assessed using a modular breath sampler (MBS, Xheal Diagnostics, Fig. 1a). Patients were instructed to inhale and exhale normally through the mouthpiece for one minute, after which they inhaled and www.nature.com/scientificreports/ exhaled deeply three times before continuing to breath normally to complete two minutes in total. The sample, collected in capture buffer, was stored at − 20 °C until further processing. Saliva was collected by instructing patients to spit in a sterile 15 or 50 mL container (Greiner) and stored at − 20 °C.

RT-qPCR.
The presence and viral load of SARS-CoV-2 were determined using RT-qPCR adapted from that of the Dutch National Institute for Public Health and the Environment (RIVM). Briefly, 500 µl material was lysed in 450 µl MagNAPure lysis/binding buffer (Roche). RNA internal extraction control (Plasmodium falciparum PfMGET ivRNA) was added prior to extraction using the MagNAPure LC Total Nucleic Acid-High Performance kits (Roche). RT-qPCR was performed using the Luna Universal Probe One-Step RTqPCR kit (NEB) with 400 nM E-gene primers (FW: 5'-ACA GGT ACG TTA ATA GTT AAT AGC GT-3' RV: 5'-ATA TTG CAG CAG TAC GCA CACA-3') and 200 nM E-gene probe (5'-FAM ACA CTA GCC ATC CTT ACT GCG CTT CG-BHQ1-3' (Biolegio)) on a CFX96 C1000 Real-Time PCR etection System (BioRad). Transcript quantities were calculated using a tenfold dilution series of E gene ivRNA. The extraction efficiency was checked in a separate RT-qPCR using the Luna Universal Probe One-Step RT-qPCR kit (NEB) with primers targeting PfMGET ivRNA.
Statistical analyses. Analyses were conducted using SPSS software, 27th version (IBM Corp., 2021) and GraphPad Prism, version 8.0.2 (GraphPad Software, 2019). Categorical data are presented as numbers with percentages and continuous data are presented as medians with interquartile ranges. Mann-Whitney U tests were used to assess differences between groups. Spearman rank correlation was used to assess the correlation between SARS-CoV-2 RNA loads in saliva and nasopharyngeal throat swabs.
Correlation of viral RNA load in saliva and nasopharyngeal throat swabs. Out of 30 patients from whom we collected saliva samples, 27 (90%) had detectable SARS-CoV-2 RNA in both saliva and the combined nasopharyngeal-throat swab. SARS-CoV-2 RNA loads in saliva and the combined nasopharyngeal throat swab were correlated (r s = 0.566, p = 0.002) (Fig. 1b).
Detection of SARS-CoV-2 RNA in exhaled air. SARS-CoV-2 RNA was detected in exhaled air in 12 (24.5%) patients (Fig. 1c) up to 23 days after symptom onset. Patients with and without detectable SARS-CoV-2 RNA in exhaled air did not differ in age, BMI, time from symptom onset to admission, length of admission, or laboratory measures ( Table 1). The median time from symptom onset to sampling was 7.5 days (IQR 6-14.3) in patients with detectable SARS-CoV-2 RNA versus 11 days (IQR 8.5-13) in patients without detectable SARS-CoV-2 RNA (p = 0.295). Interestingly, the presence of SARS-CoV-2 RNA in exhaled air was not limited to patients with high SARS-CoV-2 viral load in the combined nasopharyngeal throat swab. Furthermore, patients with and without detectable SARS-CoV-2 RNA in exhaled air showed no differences in SARS-CoV-2 RNA load in the combined nasopharyngeal throat swab (p = 0.335) or saliva (p = 0.938, Supplementary Fig. 1).

Discussion
In this study, we showed for the first time that SARS-CoV-2 RNA can be detected by RT-qPCR in exhaled air of hospitalized COVID-19 patients. SARS-CoV-2 RNA has previously been detected in air collected from COVID-19 wards 13 and in exhaled air from ambulant COVID-19 patients 14 , but not yet from hospitalized patients. In our study, sampling of exhaled air was feasible using a handheld modular breath sampler and viral RNA was detected in almost 25% of the patients.
We demonstrated that viral RNA is still detectable 7 days after disease onset in a clinical setting, later than Coleman et al. 14 who detected SARS-CoV-2 RNA after a median of 3 days of symptoms in an ambulant setting. Their positivity rate was higher, which can probably also be attributed to a longer sampling time of 30 min. It must be noted that the two-minute breathing exercise was well tolerated by all patients in our study, including patients receiving oxygen supplementation via nasal cannula as well as patients using high-flow oxygen therapy. The sampling could be prolonged to increase sensitivity, however, increased sampling time will reduce the usability of the modular breath sampler and hamper its clinical implication.
Viral load in nasopharyngeal throat swab and saliva samples was positively correlated, confirming previous observations both in inpatients with confirmed COVID-19 15 and in routine sampling 16 . Sampling saliva can be a patient-friendly alternative to nasopharyngeal throat swab sampling, but sensitivity can be an issue, as shown by the lower viral loads in saliva. www.nature.com/scientificreports/ The fact that SARS-CoV-2 RNA was measured up to 23 days after symptom onset suggests long-term persistence of viral RNA. Moreover, the absence of any association between SARS-CoV-2 positivity in exhaled air samples and viral load in nasopharyngeal throat swab and saliva samples makes contamination from the upper airway unlikely.
This study shows the feasibility of sampling exhaled air from hospitalized COVID-19 patients for the presence of SARS-CoV-2 RNA. The modular breath sampler could be used in epidemiological studies to determine the best proxy of SARS-CoV-2 contagiousness. www.nature.com/scientificreports/

Data availability
The data that support the findings of this study are available from the corresponding author, upon reasonable request.