Severe acute respiratory syndrome coronavirus 2 can be detected in exhaled aerosol sampled during a few minutes of breathing or coughing

Abstract Background The knowledge on the concentration of viral particles in exhaled breath is limited. The aim of this study was to explore if severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) can be detected in aerosol from subjects with the coronavirus disease 2019 (COVID‐19) during various types of breathing and coughing and how infection with SARS‐CoV‐2 may influence the number and size of exhaled aerosol particles. Methods We counted and collected endogenous particles in exhaled breath in subjects with COVID‐19 disease by two different impaction‐based methods, during 20 normal breaths, 10 airway opening breaths, and three coughs, respectively. Breath samples were analyzed with reverse transcription real‐time polymerase chain reaction (RT‐PCR). Results Detection of RNA in aerosol was possible in 10 out of 25 subjects. Presence of virus RNA in aerosol was mainly found in cough samples (n = 8), but also in airway opening breaths (n = 3) and in normal breaths (n = 4), with no overlap between the methods. No association between viral load in aerosol and number exhaled particles <5 μm was found. Subjects with COVID‐19 exhaled less particles than healthy controls during normal breathing and airway opening breaths (all P < 0.05), but not during cough. Conclusion SARS‐CoV‐2 RNA can be detected in exhaled aerosol, sampled during a limited number of breathing and coughing procedures. Detection in aerosol seemed independent of viral load in the upper airway swab as well as of the exhaled number of particles. The infectious potential of the amount of virus detected in aerosol needs to be further explored.


| INTRODUCTION
The coronavirus disease 2019 , caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is assumed to mainly be transmitted by respiratory droplets. However, probable aerosol transmission has been reported to occur under certain conditions. 1,2 The knowledge on the concentration of SARS-CoV-2 particles (viral load) in exhaled breath samples is limited as well as the size and concentration of exhaled endogenously generated droplets in relation to viral load. Moreover, the relation between the viral load in upper airway diagnostic samples and aerosol samples needs further clarification.
An aerosol contains micro-sized particles generated from the respiratory tract lining fluid (RTLF) in the airways. 3 The smallest fraction exhaled, that is, particles <5 μm, are mainly formed on inhalation when bronchiolar fluid film bursts. 4,5 Particle formation in small airways, occurring at normal breathing, can be increased by deep exhalation followed by deep inhalation resulting in airway closure and reopening. 5,6 The individual number and size of exhaled particles differ substantially both between and within one subject and may depend on several factors, for instance, the precise breathing pattern used. 4,7,8 It has been suggested that viruses may be enriched in exhaled particles with a diameter of <5 μm. 9 This is supported by recent findings indicating higher SARS-CoV-2 viral load in exhaled particles <5 μm compared with those >5 μm during speech and song from COVID-19 patients 10 as well as in breath samples from anesthetized SARS-CoV-2-infected macaques. 11 The PExA (particles in exhaled air) method is optimized for collection of small airway particles with a size range of approximately 0.4-5 μm, using the airway opening maneuver. 12 Another aerosolcollection device; BE (Breath Explor), also based on impaction but without particle counting and airflow data, has recently been developed. 13 Due to the design of BE where sample is collected immediately at the mouth opening, it seems highly plausible that also particles >5 μm can be sampled using this device.
The main objective of this study was to explore if SARS-CoV-2 can be detected in aerosol from subjects newly diagnosed with COVID-19, using PExA and BE. If so, we aimed to compare the viral load in exhaled aerosol collected during various types of breathing and coughing and to explore how infection with SARS-CoV-2 may influence the number and size of exhaled aerosol particles <5 μm.

| Study design and study subjects
Subjects with mild symptomatic COVID-19 disease were recruited for a first substudy (i) during September-October 2020. A second substudy (ii) took place from April to May 2021. Both studies were carried out at Sahlgrenska University Hospital, Gothenburg, Sweden.
Eligible subjects were hospital health care workers testing positive for COVID-19 according to the routine testing of hospital staff currently used at the time for the study inclusion. In the first substudy (i), diagnosis was confirmed by positive reverse transcription real-time polymerase chain reaction (RT-PCR) analysis of a combined oropharyngeal and nasopharyngeal (oro/nasopharyngeal) swab sample, and in the second substudy (ii) by a positive Rapid COVID-19 Antigen Test (CLINITEST ® ) (Siemens, Healthineers, Erlangen, Germany) obtained by nasopharyngeal swab sampling.
Breath sampling was performed (i) several days after a positive PCR of oro/nasopharyngeal sample, and (ii) immediately after antigen testing at the staff testing station. A positive antigen test was confirmed with PCR analysis of an oro/nasopharyngeal swab sample, taken the same day. All control subjects were recruited among hospital health care workers without symptoms of COVID-19 and performed a Rapid SARS-CoV-2 Antigen Test Card (Boson Biotech) (Xiamen Boson Biotech Co., Ltd, Xiamen, China) to confirm their negative COVID-19 status at the time of the breath sampling.

| Breath sampling by PExA
Characterization and collection of PExA ( Figure 1) was made using the PExA instrument (PExA AB, Gothenburg, Sweden). The principles of the PExA method have been described in detail previously by Almstrand et al. 12 In brief, the PExA instrument uses an optical particle counter (Grimm 1.108, Grimm Aerosol Technik GmbH, Ainring, Germany) and a cascade impactor for particle collection. The particle counter measures particle number concentrations and particle size in Subjects wore a nose clip and inhaled air through a highefficiency particle arresting (HEPA) filter to remove ambient particles so that only endogenous particles were exhaled into the PExA instrument.
The separate breathing procedures used for particle collection were as follows: • Normal breathing: relaxed breathing for 20 breaths.
• Airway opening maneuver: Deep exhalation to residual volume (RV), 5-s breath hold at RV followed by a rapid inhalation to total lung capacity (TLC) and finally a relaxed exhalation to RV. Repeated 10 times.
• Coughs: Deep inspiration to TLC followed by a cough. Repeated three times.
The PTFE membranes on each impactor plate with sampled particles was transferred to separate, 1.5-ml SC microtube PC-PT cryotubes (Sarsteds, Nümbrecht, Germany) and stored at À80 C, prior to RNA extraction. Particle number concentrations are expressed as n * 1,000 (kn).

| Breath sampling with BE
The handheld BE instrument (Munkplast AB, Uppsala, Sweden) was used for collection of particles in aerosol, without counting and size fractioning. The principles of the BE method have been described in detail previously by Seferaj et al. 13 Sampling was performed according to instructions by the manufacturer (VER: 2019-05-06), with minor modifications. Subjects wore a nose clip throughout the collection and performed 20 normal relaxed exhalations into the device.
Samples were then stored at À80 C, prior to RNA extraction.

| RNA extraction and detection of SARS-CoV-2 in exhaled particles
The SARS-CoV-2 PCR analysis, from which the cycle threshold

| Statistical analysis
Statistical analyses were performed using the IBM SPSS software, version 26.0 (SPSS, Chicago, IL). Nonparametric tests were used due to few subjects and skewed distributions. Mann-Whitney U test was F I G U R E 1 Schematic illustration of the particles in exhaled air (PExA) instrument set-up at collection. Subject breaths through a mouthpiece, connected to a two-way, nonre-breathing valve, where inhalation goes through a high-efficiency particle arresting (HEPA) filter and exhalation goes into the instrument. An optical particle counter samples a fraction of the exhaled air with a constant flow of 20 ml/s. The two stage inertial impactor collects particles according to size by the control of a rotary vane (RV) pump with a constant flow of 230 ml/s. A reservoir handles exhalations that exceeds the flow rate through the impactor used for comparison of continuous data and chi-square tests using Fisher's exact significance (two-sided) for comparison of categorical data between subjects with COVID-19, with and without detectable SARS-CoV-2 RNA.

| RESULTS
Ten subjects with confirmed COVID-19 were included in substudy (i), where aerosol sampling took place 8 (5-11) days after symptom onset and 3 (1-7) days after PCR-positive oro/nasopharyngeal swab. A positive aerosol sample, collected with PExA during coughing, was found in the one subject with the shortest symptom duration (5 days). On the basis of this finding, we then re-designed the protocol in order to minimize the time between symptom onset and aerosol sampling.
Twenty-five subjects with COVID-19 and with a symptom duration of 2 (0-9) days, and 11 controls, were included in the second sub- one in normal breathing and one in cough. Overall, viral load in oro/nasopharyngeal samples was high in comparison with that of samples from aerosol, and no association between viral load in upper airways and aerosol specimens was found. detectable SARS-CoV-2 RNA in the aerosol, apart from nasal congestion that was more common among those with RNA detected only in upper airway swab and not in aerosol.
Ten airway opening maneuvers produced a hundred times more particles per exhalation than that of normal breaths, as illustrated in    Aerosol sampling by impaction has been used also in previous studies. 10,11 The novelty in the present study is counting and size fractioning of particles between 0.4 and approximately 5 μm, and the use of airway opening maneuver as one of the breathing patterns studied.
Presence of virus RNA in aerosol detected with PExA was almost exclusively found on the impactor stage sampling particles <5 μm, in comparison with >5 μm, in line with previous findings 10, 11 and further supported by a recent review suggesting that SARS-CoV-2 RNA is enriched in aerosol with particle sizes <5 μm. The viral load in all aerosol samples was low in comparison with the oro/nasopharyngeal samples, which contain a larger volume of secretion, and is in line with previous studies. 10,19,20

PEER REVIEW
The peer review history for this article is available at https://publons. com/publon/10.1111/irv.12964.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.