Detection and isolation of airborne SARS‐CoV‐2 in a hospital setting

Abstract Transmission mechanisms for severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) are incompletely understood. In particular, aerosol transmission remains unclear, with viral detection in air and demonstration of its infection potential being actively investigated. To this end, we employed a novel electrostatic collector to sample air from rooms occupied by COVID‐19 patients in a major Swedish hospital. Electrostatic air sampling in conjunction with extraction‐free, reverse‐transcriptase polymerase chain reaction (hid‐RT‐PCR) enabled detection of SARS‐CoV‐2 in air from patient rooms (9/22; 41%) and adjoining anterooms (10/22; 45%). Detection with hid‐RT‐PCR was concomitant with viral RNA presence on the surface of exhaust ventilation channels in patients and anterooms more than 2 m from the COVID‐19 patient. Importantly, it was possible to detect active SARS‐CoV‐2 particles from room air, with a total of 496 plaque‐forming units (PFUs) being isolated, establishing the presence of infectious, airborne SARS‐CoV‐2 in rooms occupied by COVID‐19 patients. Our results support circulation of SARS‐CoV‐2 via aerosols and urge the revision of existing infection control frameworks to include airborne transmission.


| INTRODUC TI ON
SARS-CoV-2 is the causative agent of coronavirus disease 2019 , which has since March 23, 2022 claimed more than 6 million deaths worldwide. 1 At the onset of the COVID-19 pandemic, medical treatments such as vaccination were unavailable.
Infection prevention and control measures targeted instead the development of diagnostics and implementation of physical distancing, local and countrywide lockdowns, and disinfection protocols. Much research has since been dedicated to unfolding the routes of transmission of the virus, including the contribution of droplets, aerosols, and fomites. However, early reports emphasized droplets and fomites in transmission. 2,3 Physical-distancing guidelines to mitigate the spread of SARS-CoV-2 have advocated often vague and countryvariable "safe physical distancing" in workplaces, 4 keeping when possible, at least 6 feet (~1.8 meters) between people in healthcare facilities 5 and at least 1 meter distancing in school settings. 6 Reports on super-spreading events [7][8][9] and detection of SARS-CoV-2 RNA in hospital air [10][11][12][13][14][15] have suggested, however, that the virus may also spread through aerosols. These accounts are accumulating but are each based on limited datasets. They have also been countered by reports of negative detection of the virus in air, [16][17][18] and systematic reviews rejecting airborne transmission of the virus. 19,20 Today, the CDC and the WHO acknowledge not only the airborne route of transmission but also that the claim is not wellestablished and experimental support limited. 5,21,22 Confirmation of active SARS-CoV-2 in air remains to be thoroughly demonstrated to dispel further speculation; more studies and robust datasets are needed to establish aerosols as central to transmission of SARS-CoV-2 in the environment. A key step to establish aerosol transmission is to demonstrate the virus in air and on surfaces that cannot be explained by virus droplet deposition at that site. Such demonstration requires the combination of tactical microbiological air sampling with detection methods needed to establish the presence of airborne SARS-CoV-2. In this regard, electrostatic precipitation is emerging as a simple and efficient way to collect bioaerosols. [23][24][25][26] Herein, we used an electrostatic air sampler developed in our laboratory 27 in conjunction with an extraction-free RT-PCR for SARS-CoV-2 28 and standard viral PFU assays. We investigated the presence of airborne SARS-CoV-2 in COVID-19 patient rooms and adjoining indoor environments in a Swedish healthcare setting. Airborne investigation was performed alongside detection of the virus on out-of-reach and high-contact surfaces in the same space. THOR was placed on a tripod at 1 m from the ground and at least 2 m from the patient. Collector pieces were transferred into 0.5 ml PBS containing 0.05% Tween-80 (PBS-T)(Sigma) and vortexed for 1 min. 27 THOR devices were disinfected with 70% EtOH before moving to a new sampling location.

| Environmental sampling
Patient rooms had on average a floor area of 25 m 2 in size and were 65 m 3 in volume (Table 1). Two patients could be housed in these rooms at once. Surface samples from the room's air exhaust ventilation, the railing of the patient's bed and the floor next to the bed were collected in patient rooms prior to air sampling. The surface of the air exhaust ventilation was also swabbed in the anteroom. Surface samples were collected using MS Mini DNA/RNA buccal swabs (Isohelix, Cell Projects). Two swabs were used simultaneously on the same surface area, stored in 3 ml eNat ® preservation buffer (COPAN) and processed separately. In rooms with two occupants, each bed rail was sampled with one individual swab. The total surface area sampled at each location was ~90 cm 2 for air exhaust vents, ~100 cm 2 for bed rails and ~625 cm 2 for the floor. Indoor room air temperature, relative humidity, and CO 2 levels were measured using a pSENSE II environmental logger (Senseair AB). An illustration of THOR, the approximate location of THOR, approximate location of the environmental logger, and of surface sampling in patient and adjoining anterooms are provided ( Figure S1). The full layout of the ward has been previously reported. 29 Ten different patient rooms with adjoining anterooms were sampled in the above way. One patient room and adjoining anteroom were sampled twice but with more than 3 weeks interval between sampling. Thus, we considered our samples to include a total of 11 groups of rooms. In addition, control air sampling was done in 1 physician office and 1 physician meeting room, both located just outside the ward and empty at the time of sampling. Air sampling was also performed in 2 patient rooms and their adjoining anterooms which had been cleaned as per standard ward routines and unoccupied for at least 2 days.

Practical implications
• Support for airborne route of viral transmission by the demonstration of active SARS-CoV-2 in air from COVID-19 patient rooms and adjoining indoor environments.
• Support for aerosol transmission-based mitigating measures against SARS-CoV-2 in healthcare settings.

| Ethics statement
The study was conducted according to good clinical and scientific practices and following the ethical principles of the Declaration of

| Extraction-free RT-PCR on heat-inactivated samples (hid-RT-PCR)
Hid-RT-PCR was performed on heat-inactivated aliquots of all samples as previously described, 28 with a few modifications. Primer and probe sequences used are the same as in the original publication. 28 Briefly, after heat inactivation at 95°C for 5 min, 12 μl of the inactivated sample was added to 48 μl RT-PCR mastermix consisting of 15 μl of one-step TaqPath RT-qPCR master mix (Thermo, A15299), 4.5 μl of primer-probe mix, 1.8 μl of 10% Tween-20 solution, and nuclease-free water to achieve a final reaction volume of 60 μl.
The N1 primers and probe were used at a concentration of 500 nM and 125 nM, respectively. The thermal cycling steps were 25°C for 2 min, 50°C for 15 min, 95°C for 2 min, and 45 cycles of 95°C for 3 s and 56°C for 30 s. In RT-qPCR run, nuclease-free water was used as negative control and heat-inactivated in vitro expanded SARS-CoV-2 28,30 was used as positive control. The limit of detection for this reaction is 2-20 genome copies/µl.

| Isolation and quantification of SARS-CoV-2
Quantification of Plaque-forming units (PFUs) was performed on 90% confluent Vero E6 cells (ATCC-CRL-1586), using 6-well tissueculture plates. Briefly, environmental samples were serially diluted with DMEM and added to Vero E6 cells for 1 h at 37°C followed by removal of the inoculum media and two washes with PBS. Overlay medium consisting of 2:3 mix of 3% carboxymethyl cellulose and DMEM was added, and the plates were incubated at 37°C for 3 days.
Plates were then assessed under the microscope for cytopathic effects in line with viral-induced plaque formation and marked. For the confirmation of SARS-CoV-2 in PFUs from environmental samples, a 50 µl aliquot was obtained directly from the plaque by pipette aspiration of the semi-solid media using a wide bore 200 µl pipette tip, 50 µl of PBS were added to further solubilize the media and make it amiable to further processing. The presence of SARS-CoV-2 RNA in the plaque was determined by hid-RT-PCR.

| Data analyses
Graphical representations of data and statistical testing were performed using GraphPad Prism 9.3.1 (GraphPhad Software Inc).
Abbreviation: ND, not determined. a Median.
b Including bathroom.
TA B L E 1 Parameters of the sampled patient rooms and anterooms a 3 | RE SULTS

| Detection and distribution of SARS-CoV-2 RNA in air and on surfaces
Environmental sampling was performed in the infectious disease ward at the 940 bed, university teaching hospital, Uppsala University Hospital, Uppsala Sweden. Air and surface samples were collected from 11 rooms occupied by 15 patients with confirmed COVID-19.
One patient was asymptomatic, the remaining patients were symptomatic and in their second week of respiratory symptoms (Table 2).
A quarter of the individuals were also experiencing gastrointestinal symptoms ( Table 2). Eleven sets of samples were collected from Importantly, SARS-CoV-2 RNA was not detected in control environments (physician office and staff meeting room). Detection was also negative in two patients and adjoining anterooms cleaned after patient discharge and kept unoccupied for at least 2 days.
To investigate the impact of basic room atmospheric parameters on virus detection, we attempted to match air temperature, relative humidity, and CO 2 concentration against virus RNA Ct values from the patient room air and the surface of the air exhaust vent.
Correlation of given environmental parameters with the detection of virus RNA was, however, not observed ( Figure 2). Moreover, viral RNA Ct values in air or on investigated surfaces in the patient room did not correlate with Ct from patient upper respiratory swabs ( Figure S2).

CoV-2 in environmental samples
Despite a large return of positive air samples for SARS-CoV-2 RNA, it remained to be determined whether active virus particles could be recovered from the same air-sample material. Thus, we attempted to isolate SARS-CoV-2 from samples collected from patient and anterooms. Air samples were inoculated onto a monolayer of Vero E6 cells in in vitro culture and surprisingly plaque-forming units (PFUs) were detected from a significant number of air samples (Table 3). SARS-CoV-2 RNA could subsequently be amplified by hid-RT-PCR from approximately one-third of all detected PFUs (496/1472; 34%). Most of the SARS-CoV-2 (+) PFUs were isolated from anterooms (320/496; 65%), where also more SARS-CoV-2+ PFUs were observed (Table 3). Active SARS-CoV-2 was detected in 3 patient rooms and 8 anterooms across three different sampling days (Table 4)

| DISCUSS ION
The distinction between droplet-and aerosol-based transmissions can be semantic, as both can effectively transmit pathogens, including respiratory viruses. 31 Our study confirms the detection of A limitation of several prior studies showing virus in air has been the singular reliance of RT-PCR for virus detection. [10][11][12]15 Another has been the inability to recover PFUs from RT-PCR-positive air samples 13,14,36,37 or central exhaust filters exchanging air from COVID-19 wards. 29 In our study we used hid-RT-PCR for molecular detection of virus in air. This method has the advantage of being quick, scalable, reliant on off-the-shelf reagents, low-cost on a perreaction basis and safer due to a heat inactivation step prior to sample handling. 28 Detection of SARS-CoV-2 RNA by (extraction-free) hid-RT-PCR returned higher Ct values for air samples compared to regular extraction-based RT-qPCR but not for surface samples ( Figure S3). Nevertheless, our Ct returns are consistent with prior

Instead, aerosol-based infection control and prevention measures
should be considered also following such observations.
The higher number of PFUs recovered in our study were isolated from anterooms. This is consistent with the design intervention of the rooms in the ward, which were built to generate the lowest relative pressure in the anterooms. This enables the patient rooms to be used both for isolation of patients with airborne infectious diseases as well as for protective isolation of im- We were surprised to recover active SARS-CoV-2 from air.
Inactivation of bacteria or viruses is expected during electrostatic air sampling, 42-4 4 including on our own THOR collector. 27 Interestingly, a recent air sampling study performed in the same infectious disease unit was unable to retrieve active virus from air using a different electrostatic collector. 36 Our shorter sampling interval and ability to rapidly immerse sample into buffer may have helped preserve viral activity in our material. We speculate that an outer cuff of viral particles shielded a smaller number of virions in the interior of the aerosol from the inactivation effects of ionization, aerosol stress and/or impaction onto the collector. Indeed, impactors, impingers, or cyclone samplers are more readily used to recover active virus compared to electrostatic samples. 45 Our laboratory tests show that a high titer of SARS-CoV-2 is inactivated by almost 2 orders of magnitude after 15 min of THOR operation ( Figure S4). Thus, it is not surprising that the number of PFUs recovered from THOR air sampling was low. Given that the sampler used in this study was not adapted for recovery of active virus particles from air, the true viral load may be substantially higher than the values reported.
We did not observe a correlation between symptoms, disease onset or Ct returns from patient upper respiratory swabs and the detection of SARS-CoV-2 in air or surfaces. An early study also failed to couple symptoms with virus shedding as measured by contamination of air and surfaces in the infected patient's vicinity. 12 That said, our patients were entering their second week of disease, and F I G U R E 2 Environmental parameters do not correlate with detection of SARS-CoV-2 in air. Environmental parameters were recorded in patient rooms before air sampling. Ct values obtained from air samples and the air exhaust vent surface swabs by hid-RT-PCR are plotted against room air temperature (A and B), relative humidity (C and D), and CO 2 concentration (E and F). Continuous line represents the best fit regression curve (3rd order polynomial) and dashed lines the 95% confidence intervals. No correlations were observed for the given parameters.