1 The removal of airborne SARS-CoV-2 and other microbial bioaerosols by air filtration on COVID-19 surge units Andrew Conway Morris PhD± 1,2, Katherine Sharrocks DPhil± 3, Rachel Bousfield BMBS ± 3,4, Leanne Kermack MSc± 5, Mailis Maes MPhil 5, Ellen Higginson PhD 5, Sally Forrest BSc 5, Joana Pereira- Dias MSc 5, Claire Cormie5, Tim Old MA3, Sophie Brooks MSc3, Islam Hamed FRCA1, Alicia Koenig MBChB 1, Andrew Turner6, Paul White PhD6,7, R. Andres Floto FRCP8,9, Gordon Dougan D.Sc5, Effrossyni Gkrania-Klotsas PhD ¥ 3, Theodore Gouliouris PhD¥ 3,4, Stephen Baker PhD5, and Vilas Navapurkar FFICM1* 1 The John Farman ICU, Cambridge University Hospitals NHS Foundation Trust, Hills Rd, Cambridge CB2 0QQ, UK 2 University Division of Anaesthesia, Department of Medicine, University of Cambridge School of Clinical Medicine, Cambridge Biomedical Campus, Cambridge, UK 3 Department of Infectious Diseases, Cambridge University Hospitals NHS Foundation Trust, Hills Rd, Cambridge CB2 0QQ, UK 4 Clinical Microbiology Laboratory, Cambridge University Hospitals NHS Foundation Trust, Hills Rd, Cambridge CB2 0QQ, UK 5 Cambridge Institute of Therapeutic Immunology and Infectious Disease, Department of Medicine, University of Cambridge School of Clinical Medicine, Cambridge Biomedical Campus, Cambridge, UK 6 Department of Clinical Engineering, Cambridge University Hospitals NHS Foundation Trust, Cambridge UK 7 Medical Technology Research Centre and School of Medicine, Anglia Ruskin University, Chelmsford, UK 8 Molecular Immunity Unit, University of Cambridge Department of Medicine, MRC-Laboratory of Molecular Biology, Cambridge, UK 2 9 Cambridge Centre for Lung Infection, Royal Papworth Hospital, Cambridge UK ± The following authors contributed equally to this work ¥ The following authors contributed equally to this work * Address for Correspondence Vilas Navapurkar The John Farman ICU, Addenbrooke’s Hospital, Hills Rd, Cambridge CB2 0QQ, UK Vilas.navapurkar@addenbrookes.nhs.uk Alternative contact for correspondence Andrew Conway Morris, University Division of Anaesthesia, Box 94, Level 4, Addenbrooke’s Hospital, Hills Rd, Cambridge CB2 0QQ, UK ac926@cam.ac.uk Key words SARS-CoV-2, air filtration, COVID-19, nosocomial infection, airborne pathogens Running head Removal of airborne SARS-CoV-2 Word count: Summary 49 main text 1499 Key words SARS-CoV-2; COVID-19; hospital infection control; air filtration 3 Summary Airborne SARS-CoV-2 was detected in a COVID-19 ward before activation of portable HEPA-air filtration, but not during the week of filter operation; SARS-CoV-2 was again detected when the filter was off. Airborne SARS-CoV-2 was infrequently detected in a COVID-19 ICU. Filtration significantly reduced other microbial bioaerosols in both settings. 4 Introduction 1 Airborne dissemination is likely an important transmission route for SARS-CoV-2[1], with SARS-2 CoV-2 RNA detected in air samples from COVID-19 wards[1,2]. Despite the use of personal 3 protective equipment (PPE), there are multiple reports of patient-to-healthcare worker transmission of 4 SARS-CoV-2[3], potentially through the inhalation of viral particles[4] . There is a need to improve 5 the safety for healthcare workers and patients by decreasing airborne transmission of SARS-CoV-6 2[4]. Portable air filtration systems, that combine high efficiency particulate filtration and ultraviolet 7 (UV) light sterilisation, may be a scalable solution for removing respirable SARS-CoV-2[5]. A recent 8 review by the UK Scientific Advisory Group for Emergencies modelling group found limited data 9 regarding the effectiveness of such devices[6]. Here we present the first data providing evidence for 10 the removal of SARS-CoV-2 and microbial bioaerosols from the air using portable air filters with UV 11 sterilisation on a COVID-19 ward. 12 13 Methods 14 The study was conducted in two repurposed COVID-19 units in Addenbrooke’s Hospital, Cambridge, 15 UK. One area was a ‘surge ward’ (ward) managing patients requiring simple oxygen therapy or no 16 respiratory support, the second was a ‘surge ICU’ (ICU) managing patients requiring invasive and 17 non-invasive (non-invasive ventilation, high flow nasal oxygen) respiratory support. The ward was a 18 fully occupied four-bedded bay (top left panel Fig. 1A). The ICU was fully occupied five-bedded bay, 19 with a supra-capacity sixth occupied bed used in week 2 (top left panel Fig. 1B). Both units were 20 passively ventilated, with 2-4 air-changes per hour at baseline. 21 22 In the ward we installed an AC1500 HEPA14/UV steriliser (Filtrex, Harlow, UK), in the ICU we 23 installed a Medi 10 HEPA13/UV steriliser (Max Vac, Zurich, Switzerland). The air filters were placed 24 in fixed positions before the initiation of the three-week study period (Fig. 1A/B), switched on at the 25 beginning of week two and run continuously from Sunday to Sunday for 24 hours per day, providing 26 approximately 5-10 room-volume filtrations per hour. 27 5 We performed a crossover evaluation, with the primary outcome being detection of SARS-CoV-2 28 RNA in the various size fractions of air samples. Air sampling was conducted using National Institute 29 for Occupational Safety and Health (NIOSH) BC 251 two-stage cyclone aerosol samplers[7] (B 30 Lindsley, CDC), operated in accordance with previous studies [7,8]. Air samplers were assembled 31 daily with a control sampler left in a sealed bag. Samplers were placed adjacent to the air filter inlet 32 and the other at approximately four meters from the filter and no closer than two meters to patients. 33 In ICU two distant samplers were used, one mounted at head height and one at bed height. Samplers 34 were operated on weekdays (0815hrs to 1415hrs) for three consecutive weeks. After sampling, 35 samplers were disassembled using sterile technique. The samples were processed then stored at −80°C 36 until analysis 37 38 Nucleic acids were extracted from each NIOSH sampler component (tubes containing large aerosols, 39 medium aerosols, and filter). Methodological details including extractions, RT-qPCR for SARS-CoV-40 2 and high throughput qPCR assays for a range of bacterial, viral, and fungal pathogens are in the 41 supplemental methods (organisms listed in supplemental table 1). Differences in numbers of 42 pathogens detected with filters on and off were compared by Mann-Whitney U test, p ≤0.05 was 43 considered significant. 44 45 Results 46 From January 18th to February 5th beds in the ward and ICU were at 100% occupancy; 15 patients 47 admitted to the ward and 14 admitted to the ICU over the sampling period. All patients were 48 symptomatic and tested positive for SARS-CoV-2 RNA. 49 50 In the ward, during the first week whilst the air filter was inactive, we were able to detect SARS-CoV-51 2 on all sampling days; RNA was detected in both the medium (1-4µM) and the large (>4µM) 52 particulate fractions (lower panel Fig. 1A). SARS-CoV-2 RNA was not detected in the small (<1μM) 53 particulate filter. The air filter was run continuously in week 2; we were unable to detect SARS-CoV-54 2 RNA in any of the sampling fractions on any of the five testing days. We completed the study by 55 6 repeating the sampling with an inactive air filter. As in week one, we were able to detect SARS-CoV-56 2 RNA in the medium and the large particulate fractions on 3/5 days of sampling (a sample without 57 tube size indicated tested positive on day 5) (lower panel Fig. 1A). SARS-CoV-2 RNA was not 58 detected from the control sampler. 59 60 We subjected the extracted nucleic acid preparations to high-throughput qPCR to detect a range of 61 viral, bacterial, and fungal targets. In week one, we detected nucleic acid from multiple viral, 62 bacterial, and fungal pathogens on all sampling days (top middle and right panels Fig. 1A). In 63 contrast, when the air filter was switched on, we detected yeast targets only on a single day, with a 64 significant reduction (p=0.05) in microbial bioaerosols when the air filter was operational (Fig. 1A). 65 Using this high-throughput approach, SARS-CoV-2 RNA was detected on 4/5 days tested in week 1 66 but was again absent in week 2. We were unable to generate multiplex data for week three due to 67 sample degradation following SARS-CoV-2 RNA amplification. 68 69 In contrast to the ward, we found limited evidence of airborne SARS-CoV-2 in weeks one and three 70 (filter off) but detected SARS-CoV-2 RNA in a single sample in the medium (1-4µM) particulates on 71 week 2 (filter on) (lower panel Fig. 1B). This contrary result did not reflect a general lack of 72 bioaerosols in the ICU, which demonstrated a comparable quantity and array of pathogen associated 73 nucleic acids to that seen in the unfiltered ward air on week one (top middle and right panels Fig. 1B). 74 Again, the use of the air filtration device significantly (p=0.05) reduced the microbial bioaerosols (Fig 75 1B); with only three organism types detected on two of the sampling days. SARS-CoV-2 RNA was 76 only detected once during week one on the high throughput qPCR assay. 77 78 Discussion 79 Our study represents the first report of removal of airborne SARS-CoV-2 in a hospital environment 80 using combined air filtration and UV sterilisation technology. Specifically, we provide evidence for 81 the circulation of SARS-CoV-2 in a ward within airborne droplets of >1µM. Droplets of 1-4µM are 82 7 likely a key vehicle for SARS-CoV-2 transmission[9], as they remain airborne for a prolonged period 83 and can deposit in the distal airways. Recent data has shown that exertional respiratory activity, such 84 as that seen in patients with COVID-19, increases the release of 1-4 µM respiratory aerosols, relative 85 to conventionally defined ‘aerosol generating procedures’ such as non-invasive respiratory support 86 [10]. Patients in ICU are commonly at a later stage of disease, and may shed less virus as a result. 87 These data are consistent with our observations, suggesting that aerosol precautions may be more 88 important in conventional wards than in well defined ‘aerosol risk areas’. 89 90 The sampling and detection of airborne viruses poses several technological challenges, and there 91 remains no agreed standard for their use or interpretation[11]. However, the detection of SARS-CoV-92 2 RNA by RT-qPCR (albeit at a high CT value), and the lack of detection during use of an air 93 sterilisation system, adds to a growing body of evidence implicating the airborne transmission of 94 SARS-CoV-2[1]. The detection of SARS-CoV-2 RNA in the air of a ward managing patients with 95 COVID-19 intimates that this is a key mechanism by which healthcare professionals could become 96 infected. The removal of airborne viral particles and other pathogens may help reduce the likelihood 97 of hospital-acquired respiratory infections. This reduction may be by both decreasing the load of 98 respirable particles and by removal of larger droplets that can facilitate fomite-associated spread[11]. 99 The clearance of bioaerosol was not restricted to SARS-CoV-2. Although the impact of air filtration 100 on nosocomial infection is uncertain[5,12], the broad range of pathogens removed in this study 101 suggests potential for benefit beyond SARS-CoV-2. 102 103 This study has limitations. The evaluation was conducted in two rooms and there are no data defining 104 the optimal air changes required to remove detectable pathogens with the specified devices, nor their 105 impact in better ventilated facilities. Given the large volume of air within the room and the stability of 106 viruses in the sampling fluid, it was predictable that the amount of SARS-CoV-2 detected would be 107 minimal. However, negative results from the control sampler, and the striking but reversible effect of 108 the air filtration devices, suggest these are not false positive detections and we cannot exclude the risk 109 8 of airborne infection. Future studies should examine whether air filtration devices have an impact on 110 healthcare professional and patient focussed outcomes, including measuring infection/exposure as an 111 endpoint, as well as assessing potential harm, such as noise, reduced ambient humidity or impact on 112 delivery of care. 113 114 We were able to detect airborne SARS-CoV-2 RNA in a repurposed COVID-19 ‘surge ward’ and 115 found that air filtration can remove SARS-CoV-2 RNA below the limit of qPCR detection. SARS-116 CoV-2 was infrequently detected in the air of a ‘surge ICU’; however, the device retained its ability to 117 reduce microbial bioaerosols. Portable air filtration devices may mitigate the reduced availability of 118 airborne infection isolation facilities when surges of COVID-19 patients overwhelm healthcare 119 resources and improve safety of those at risk of exposure to respiratory pathogens such as SARS-120 CoV-2. 121 122 123 Author contributions 124 ACM conceptualisation, methodology, data analysis, writing-original draft 125 KS investigation, supervision, writing-review and editing 126 RB investigation, supervision, writing-review and editing 127 LK investigation, data analysis, supervision, writing-review and editing 128 MM investigation, writing-review and editing 129 EH investigation, data analysis, writing-review and editing 130 SF investigation, writing-review and editing 131 JD investigation, writing-review and editing 132 TO investigation, writing-review and editing 133 SBr investigation, writing-review and editing 134 IH investigation, writing-review and editing 135 AK investigation, writing-review and editing 136 AT investigation, writing-review and editing 137 9 PW conceptualisation, provision of resources, writing-review and editing 138 AF Provision of resources, writing-review and editing 139 GD conceptualisation, provision of resources, supervision, writing-review and editing 140 EG conceptualisation, supervision, investigation, writing-review and editing 141 TG conceptualisation, supervision, investigation, writing-review and editing 142 SB conceptualisation, provision of resources, supervision, data analysis, writing-original draft 143 VN conceptualisation, provision of resources, supervision, data analysis, writing-original draft 144 145 Funding 146 This work was supported by a Wellcome senior research fellowship to Stephen Baker 147 (215515/Z/19/Z), and NIHR AMR Research Capital Funding Scheme [NIHR 200640]. Andrew 148 Conway Morris is supported by a Clinician Scientist Fellowship from the Medical Research Council 149 (MR/V006118/1). Mailis Maes and Sally Forrest are funded by the National Institute for Health 150 Research [Cambridge Biomedical Research Centre at the Cambridge University Hospitals NHS 151 Foundation Trust]. The views expressed are those of the authors and not necessarily those of the 152 NHS, the NIHR or the Department of Health and Social Care. The funders had no role in the design 153 and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, 154 review, or approval of the manuscript; and decision to submit the manuscript for publication. 155 156 The study was registered as a service evaluation with Cambridge University Hospitals NHS 157 Foundation Trust (Service Evaluation Number PRN 9798). 158 159 Data sharing statement 160 Data for organisms detected by single plex PCR (SARS-CoV-2) and high throughput PCR (bioaersol) 161 are included as supplemental data files. 162 163 Conflicts of interest 164 10 Vilas Navapurkar is the founder, Director, and shareholder of Cambridge Infection Diagnostics Ltd. 165 Andrew Conway-Morris, Paul White, Gordon Dougan and Stephen Baker are members of the 166 Scientific Advisory Board of Cambridge Infection Diagnostics Ltd. Theodore Gouliouris has 167 received a research grant from Shionogi. R Andres Floto has received research grants and/or 168 consultancy payments from GSK, AZ, Chiesi, Shionogi, Insmed, Thirty Technology. Effrossyni 169 Gkrania-Klotsas has received a National Institute of Health Research Greenshoots Award 170 171 References 172 1. Greenhalgh T, Jimenez JL, Prather KA, Tufekci Z, Fisman D, Schooley R. Ten scientific reasons in 173 support of airborne transmission of SARS-CoV-2. Lancet 2021; 397:1603–1605. 174 2. Zhou J, Otter JA, Price JR, et al. Investigating SARS-CoV-2 surface and air contamination in an 175 acute healthcare setting during the peak of the COVID-19 pandemic in London. Clin Infect Dis 2020; 176 :ciaa905-. 177 3. Illingworth CJ, Hamilton WL, Warne B, et al. Superspreaders drive the largest outbreaks of 178 hospital onset COVID-19 infections. Elife 2021; 10:e67308. 179 4. Fennelly KP. Particle sizes of infectious aerosols: implications for infection control. Lancet Respir 180 Medicine 2020; 8:914–924. 181 5. Morawska L, Allen J, Bahnfleth W, et al. A paradigm shift to combat indoor respiratory infection. 182 Science 2021; 372:689–691. 183 6. UK scientific advisory group for emergencies. Potential application of air cleaning devices and 184 personal decontamination to manage transmission of COVID-19, 4 November 2020 185 (https://www.gov.uk/government/publications/emg-potential-application-of-air-cleaning-devices-and-186 personal-decontamination-to-manage-transmission-of-covid-19-4-november-2020 accessed 187 18/9/2021) 188 11 7. Lindsley WG, Schmechel D, Chen BT. A two-stage cyclone using microcentrifuge tubes for 189 personal bioaerosol sampling. J Environ Monitor 2006; 8:1136–1142. 190 8. Killingley B, Greatorex J, Digard P, et al. The environmental deposition of influenza virus from 191 patients infected with influenza A(H1N1)pdm09: Implications for infection prevention and control. J 192 Infect Public Heal 2016; 9:278–288. 193 9. Fears AC, Klimstra WB, Duprex P, et al. Early Release - Persistence of Severe Acute Respiratory 194 Syndrome Coronavirus 2 in Aerosol Suspensions - Volume 26, Number 9—September 2020 - 195 Emerging Infectious Diseases journal - CDC. Emerg Infect Dis 2001; 26:2168–2171. 196 10. Wilson NM, Marks GB, Eckhardt A, et al. The effect of respiratory activity, non‐invasive 197 respiratory support and facemasks on aerosol generation and its relevance to COVID‐19. Anaesthesia 198 2021; 199 11. Stockwell RE, Ballard EL, O’Rourke P, Knibbs LD, Morawska L, Bell SC. Indoor hospital air and 200 the impact of ventilation on bioaerosols: a systematic review. J Hosp Infect 2019; 103:175–184. 201 12. Mousavi ES, Kananizadeh N, Martinello RA, Sherman JD. COVID-19 Outbreak and Hospital Air 202 Quality: A Systematic Review of Evidence on Air Filtration and Recirculation. Environ Sci Technol 203 2021; 55:4134–4147. 204 205 206 207 208 209 210 211 212 12 213 214 Figure 1. Bioaerosol detection in specific air sampler fractions over the three-week testing period on 215 a ‘surge’ ward and ‘surge’ ICU. 216 A) Data from ‘surge’ ward. Panels depict; top left: Layout of the room on the ‘surge’ ward with four 217 beds. The air filter was installed in the marked location and set to operate at 1,000 m3/hour with a 218 room volume of approximately 107 m3. Top middle: Stacked bar chart showing collated total number 219 of bioaerosol detections during weeks one (filter off) and two (filter on) *p=0.05 by Mann-Whitney U 220 test. Top right: CT values of detected pathogens by high-throughput qPCR when filter switch on and 221 off. Bottom: CT values for the single qPCR SARS-CoV-2 detection when filter switch on and off. 222 B) Data from ‘surge’ ICU. Panels depict; top left; Layout on the ‘surge’ ICU with six beds including 223 the addition of a further supra-capacity bed to increase occupancy (labelled with red box). The air 224 filter was installed in the marked location and set to operate at 1,000 m3/hour with a room volume of 225 approximately 195m3. Top middle: Stacked bar chart showing collated total number of bioaerosol 226 detections during weeks one (filter off) and two (filter on) *p=0.05 by Mann-Whitney U test. Top 227 right: CT values of detected pathogens by high-throughput qPCR when filter switch on and off. 228 Bottom: CT values for the single qPCR SARS-CoV-2 detection when filter switch on and off. N.B. 229 variation in CT values is a function of the microfluidics technology, and do not reflect higher 230 bioaerosol burdens. 231 232 233 234 18 Supplemental methods for “The removal of airborne SARS-CoV-2 and other microbial 239 bioaerosols by air filtration on COVID-19 surge units” 240 Setting 241 The study was conducted in two repurposed COVID-19 units in Addenbrooke’s Hospital, Cambridge, 242 UK in January/February 2021 when the alpha variant (lineage B1.1.7) comprised >80% of circulating 243 SARS-CoV-2 S1. 244 245 Air changes in wards 246 Both the room in the ‘surge’ ward and ‘surge’ ICU were passively ventilated, without forced air 247 changes. 248 249 Air filtration devices 250 The devices used were a AC1500 HEPA14/UV steriliser (Filtrex, Harlow, UK), whilst in the ICU we 251 installed a Medi 10 HEPA13/UV steriliser (Max Vac, Zurich, Switzerland). The filter system has 252 three stage particulate system: a coarse panel pre-filter, a secondary V-flow filter (ePM1=80%), and a 253 HEPA filter, tested to EN1822 standards and >99.99% efficient at removing 0.3-micron particles. The 254 filters are consistently exposed to 253nm UV-C lamps, certified to be 100% effective in removing 255 microbiological agents. The units are certified to supply ISO5-EN ISO 14644 Cleanroom standard air 256 (Class 100 US FED 209E). As the devices do not meet medical device electrical safety standards 257 (EN60601) they were operated at a distance of ≥1.5metres from any patient. 258 259 National Institute for Occupational Safety and Health (NIOSH) BC 251 two-stage cyclone aerosol 260 samplers 261 Each sampler collects large (>4 μM) particles into a 15 mL centrifuge tube, medium (1–4 μM) 262 particles into a 1.5 mL centrifuge tube, and small (<1 μM) particles in a 37-mm diameter, 263 polytetrafluoroethylene filter with 3-μm poresS2. Once sampling was complete samplers were 264 disassembled using sterile technique and the filter papers were transferred to 15 ml Falcon tubes . The 265 19 pump flow rate was set at 3.5 L of air min−1, using a flow calibrator and sampling duration set at six 266 hours (collecting a total of 1,260 L/day), following criteria from previous studies demonstrating the 267 capture of airborne viruses for RT-PCR detectionS3-7. 268 269 Nucleic acid extraction and polymerase chain reactions (PCR) 270 To facilitate solubilisation of nucleic acids, tubes were left on a tube rotator overnight at 4°C in lysis 271 buffer containing 4M guanidine thiocyanate and 0.5% β-mercaptoethanol. After overnight 272 solubilisation, all lysis buffer was removed from tubes and the extraction completed as described by 273 Sridhar et alS8. All samples were eluted in 100 µl nuclease-free water and stored at -80°C until 274 required for qPCR. 275 276 SARS-CoV-2 PCR 277 SARS-CoV-2 was detected in samples using the primers and method described previouslyS8 . Briefly, 278 5 µl of the nucleic acid extract was combined with 20 µl master mix (12.5 µl 2X Luna Universal 279 Probe One-Step reaction mix, 0.5 µl Wu forward and reverse primers (20 pmoles/µl), 0.3 µl Wu 280 FAM-MGB probe (10 pmoles/µl), 0.5 µl MS2 forward and reverse primers (10 pmoles/μl), 0.3 µl 281 MS2 ROX probe (10 pmoles/µl), 1 µl Luna WarmStart RT Enzyme Mix (New England Biolabs, 282 Hitchin, UK) and 3.9 µl nuclease-free water) in a 96-well plate. Reactions were then run on the 283 QuantStudio 7 Flex real-time PCR system (Thermofisher Scientific, Waltham, MA, USA) using the 284 following cycling conditions: 2 minutes at 25°C, 15 minutes at 50°C (reverse transcription), 2 minutes 285 at 90°C and then 45 cycles of 3 seconds at 95°C followed by 30 seconds at 60°C. 286 287 Bioaerosol high-throughput qPCR 288 Other pathogens were detected using a BioMark HD qPCR system (Fluidigm, Cambridge, UK). To 289 prepare individual 10X assays for the BioMark HD qPCR, 2.5 µl of each forward and reverse primer 290 pair (100 µM), was combined with 25 µl of 2X Assay Loading Reagent and 22.5 µl of TE buffer to a 291 final primer concentration of 500nM. Microbial targets are listed in Table S1. Pooled assays for pre-292 amplification were produced by combining 1µl of all primer pairs and diluting to a final volume of 293 20 200 µl in TE buffer (Invitrogen, Thermofisher Scientific) to give a final primer concentration of 294 500 nM. Stock solutions of the pooled and individual assay mixtures were stored at −20°C. 295 296 4µl of nucleic acid extract was first reverse-transcribed using Fluidigm Reverse Transcriptase as per 297 manufacturer instructions. Pre-amplification of cDNA was then performed to minimise sampling bias, 298 using the Fluidigm PreAmp Master Mix Kit. 1.25 µl of reverse transcribed samples were then 299 combined with 2.5 µl 2X PreAmp Master Mix, 0.5 µl pooled primers (500nM), 0.75 µl and nuclease-300 free water. Reactions were then run using cycling conditions of 95 °C for 10 minutes, followed by 17 301 cycles of 95 °C for 15 seconds and 60 °C for 2 minutes, and a final hold at 4 °C. Finally, samples 302 underwent exonuclease I (Exo-I) (NEB) treatment to degrade any remaining single stranded DNA in 303 accordance with manufacturer’s instructions, before dilution 1:5 with TE buffer. 304 305 Samples were prepared for IFC (integrated fluidics circuit) loading as per manufacturer’s instructions, 306 with 2.5 µl of 2× SsoFast™ EvaGreen® Supermix Low ROX (BioRad, Watford, UK) and 0.25 µl of 307 20× DNA Binding Dye Sample Loading Reagent combined with 2.25 µl of the Exo-I treated samples. 308 5 µl of each assay mix (see above) and sample mix was loaded into the suitable IFC inlets and the IFC 309 was loaded using the Fluidigm Juno. Once complete, the IFC was moved to the BioMark HD for 310 qPCR using the pre-programmed thermal protocol: GE Fast 96x96 PCR+Melt v2.pcl. 311 Preliminary thresholding of the amplification data was completed using the Fluidigm Real-Time PCR 312 Analysis Software, before raw data was exported to R (RStudio, Boston, USA) to apply manually 313 defined melting curve peak thresholds. Positive samples were determined to be those with Ct values 314 <= 23 and with melt curves within the previously determined range for that assay target. 315 316 Statistical analyses 317 Differences in the number of pathogens detected when air filter was on and off were compared by 318 Mann-Whitney U-test. Statistical significance was inferred when p values were ≤0.05. Statistical 319 testing and graphs generation were conducted in R studio. 320 21 321 Supplemental Table 1. Bacterial, fungal, and viral targets which formed the targets of the microbial 322 bioaerosol high-thoughput qPCR*. 323 324 Bacteria Mycobacteria Atypical bacteria Fungi Viruses Acinetobacter baumannii Mycobacterium tuberculosis Chlamydia pneumoniae Aspergillus fumigatus Adenovirus Bordetella pertussis Mycobacterium spp Chlamydia psittaci Aspergillus spp Bocavirus Bordetella parapertussis Coxiella burnetii Candida spp HCoV 229E Citrobacter spp Legionella pneumophila Fungal ribosomal 18S HCoV NL63 Corynebacterium diphtheriae Legionella spp HCoV OC43 Escherichia coli Mycoplasma pneumoniae HCoV HKU1 Enterococcus faecium Leptospira spp Cytomegalovirus Enterococcus faecalis Epstein-Barr virus Enterococcus sp Enterovirus Elizabethkingia meningoseptica Herpes Simplex virus Haemophilus influenzae Influenza A virus Klebsiella pneumoniae Influenza B virus Moraxella catarrhalis Human Metapneumovirus Morganella morganii Measles morbillivirus Neisseria meningitidis Mumps virus Proteus mirabilis Parainfluenza Pseudomonas aeruginosa Parechovirus Serratia marcescens Rhinovirus Staphylococcus aureus Respiratory syncytial virus Staphylococcus epidermidis Rubella virus Coagulase negative staphylococci SARS-CoV-2 Stenotrophomonas maltophilia Varicella zoster virus Streptococcus pneumoniae Streptococcus pyogenes 325 *Species were selected for their known respiratory pathogenicity or frequency as agents of hospital-326 acquired infection. HCoV human corona virus, SARS-CoV-2 severe acute respiratory syndrome 327 coronavirus 2. Loading control was with bacteriophage MS2. (Primer sequences available on request) 328 329 22 330 Data availability. 331 qPCR and high throughput PCR results are contained as supplemental spreadsheets labelled 332 SARS_AIR_qPCR and Fluidigm_Air_Raw1 respectively. A data dictionary is included in the 333 supplemental section below. 334 335 Data dictionary 336 File: Fluidigm_Air_Raw1 337 File refers to high-throughput PCR obtained from Biomark HD device 338 Sample.Name – sample identifier: unit, sample number, day 339 Day – day of sampling 340 Day_number -day of sampling as number 341 Filter_location -near: close to air filter, away: away from filter, away_1: away from filter 342 (bed height, ICU only), away_1.7:away from filter(head height, ICU only). 343 Week -week of evaluation (1, 2 or 3) 344 Filter_status -off:air filter present but not operational, on:air filter present and operational. 345 Unit -location of sampler: Ward: ward, ICU: ICU, Control:sampler assembled and placed in 346 sealed bag. 347 Aerosol_Fraction- Large (>4µM), medium (1-4µM), small (<1µM) 348 Ct.Value-Cycles to threshold value 349 Pathogen- name of pathogen identified 350 Classification- type of pathogen identified 351 Interpretation- positive:appropriate melt dynamics, negative:inappropriate melt dynamics 352 (where Ct and pathogen indicated) or nothing detected, failed:failure of internal QC 353 354 355 File: SARS_AIR_qPCR 356 357 Sample.Name – sample identifier: unit, sample number, day 358 Day_number -day of sampling as number 359 Unit -location of sampler: Ward: ward, ICU: ICU Control:sampler assembled and placed in 360 sealed bag. 361 Filter_Location-off:air filter present but not operational, on:air filter present and operational. 362 Aerosol_Fraction- Large (>4µM), medium (1-4µM), small (<1µM) 363 Ct.Value-Cycles to threshold value 364 Week- week of evaluation (1, 2 or 3) 365 Filter status-off:air filter present but not operational, on:air filter present and operational. 366 Interpretation- positive:appropriate melt dynamics, negative:inappropriate melt dynamics 367 (where Ct and pathogen indicated) or nothing detected, failed:failure of internal QC 368 369 370 23 371 372 373 374 Supplemental references 375 S1. Meredith LW, Hamilton WL, Warne B, et al. Rapid implementation of SARS-CoV-2 376 sequencing to investigate cases of health-care associated COVID-19: a prospective genomic 377 surveillance study. Lancet Infect Dis. 2020 ;20:1263-1271 378 379 S2. Lindsley WG, Schmechel D, Chen BT. A two-stage cyclone using microcentrifuge tubes for 380 personal bioaerosol sampling. J Environ Monit. 2006 ;8:1136-42. 381 382 S3. Chia PY, Coleman KK, Tan YK, et al. Novel Coronavirus Outbreak Research Team. Detection 383 of air and surface contamination by SARS-CoV-2 in hospital rooms of infected patients. Nat 384 Commun. 2020;11(1):2800. 385 386 S4. Killingley B, Greatorex J, Digard P, et al The environmental deposition of influenza virus 387 from patients infected with influenza A(H1N1)pdm09: Implications for infection prevention 388 and control. J Infect Public Health. 2016;9:278-88. 389 390 S5. Coleman KK, Sigler W V. Airborne Influenza A Virus Exposure in an Elementary School. Sci 391 Rep. 2020 Feb 5;10:1859. 392 393 S6. Blachere FM, Lindsley WG, Slaven JE, et al. Bioaerosol sampling for the detection of 394 aerosolized influenza virus. Influenza Other Respir Viruses. 2007;1:113-20. 395 396 S7. Blachere FM, Lindsley WG, Pearce TA, et al. Measurement of airborne influenza virus in a 397 hospital emergency department. Clin Infect Dis. 2009;48:438-40. 398 399 S8. Sridhar S, Forrest S, Kean I, et al. A blueprint for the implementation of a validated approach 400 for the detection of SARS-Cov2 in clinical samples in academic facilities. Wellcome Open 401 Res. 2020;5:110 402