key: cord-0981618-loi90cje
authors: Kotwa, Jonathon D; Jamal, Alainna J; Mbareche, Hamza; Yip, Lily; Aftanas, Patryk; Barati, Shiva; Bell, Natalie G; Bryce, Elizabeth; Coomes, Eric; Crowl, Gloria; Duchaine, Caroline; Faheem, Amna; Farooqi, Lubna; Hiebert, Ryan; Katz, Kevin; Khan, Saman; Kozak, Robert; Li, Angel X; Mistry, Henna P; Mozafarihashjin, Mohammad; Nasir, Jalees A; Nirmalarajah, Kuganya; Panousis, Emily M; Paterson, Aimee; Plenderleith, Simon; Powis, Jeff; Prost, Karren; Schryer, Renée; Taylor, Maureen; Veillette, Marc; Wong, Titus; Zhong, Xi Zoe; Mc Arthur, Andrew G; Mc Geer, Allison J; Mubareka, Samira
title: Surface and air contamination with SARS-CoV-2 from hospitalized COVID-19 patients in Toronto, Canada, March-May 2020
date: 2021-11-27
journal: J Infect Dis
DOI: 10.1093/infdis/jiab578
sha: 8e45f4b67aedef140aed883c3d7345e00af42be9
doc_id: 981618
cord_uid: loi90cje

BACKGROUND: We determined the burden of SARS-CoV-2 in air and on surfaces in rooms of patients hospitalized with COVID-19 and investigated patient characteristics associated with SARS-CoV-2 environmental contamination. METHODS: Nasopharyngeal swabs, surface, and air samples were collected from the rooms of 78 inpatients with COVID-19 at six acute care hospitals in Toronto from March to May 2020. Samples were tested for SARS-CoV-2 RNA, cultured to determine potential infectivity, and whole viral genomes were sequenced. Association between patient factors and detection of SARS-CoV-2 RNA in surface samples were investigated. RESULTS: SARS-CoV-2 RNA was detected from surfaces (125/474 samples; 42/78 patients) and air (3/146 samples; 3/45 patients); 17% (6/36) of surface samples from three patients yielded viable virus. Viral sequences from nasopharyngeal and surface samples clustered by patient. Multivariable analysis indicated hypoxia at admission, PCR-positive nasopharyngeal swab (cycle threshold of ≤30) on or after surface sampling date, higher Charlson co-morbidity score, and shorter time from onset of illness to sampling date were significantly associated with detection of SARS-CoV-2 RNA in surface samples. CONCLUSIONS: The infrequent recovery of infectious SARS-CoV-2 virus from the environment suggests that the risk to healthcare workers from air and near-patient surfaces in acute care hospital wards is likely limited.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged in December 2019 causing the coronavirus disease 2019 (COVID-19) pandemic [1] and many hospital outbreaks of COVID-19 [2] . Understanding the role of surface and air (environmental) contamination in the transmission of SARS-CoV-2 is essential to ensuring the prevention of transmission of SARS-CoV-2 between patients and to healthcare workers in acute care hospitals.

SARS-CoV-2 RNA has been detected from surfaces and air in hospitals [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] . However, a minority of studies have attempted to culture virus [13, 14] . This limits our understanding of exposure and transmission risk.

This study aimed to determine the burden of SARS-CoV-2 in the air and on surfaces in hospital rooms of acutely ill inpatients with COVID-19 in Toronto, Ontario, Canada. We determined the association between patient factors and detection of SARS-CoV-2 from environmental samples. We also apply a genomics approach to environmental samples for SARS-CoV-2, thus linking environmental contamination of SARS-CoV-2 to the source [15] [16] [17] [18] .

The Toronto Invasive Bacterial Diseases Network (TIBDN) performs population-based surveillance for infectious diseases in metropolitan Toronto and the regional Municipality of Peel, south-central [19] .

Demographic, clinical, and COVID-19 risk factor data were collected by participant interview and chart review. Study staff obtained NP swabs from patients at enrollment and every three days until refusal, hospital discharge, or death [20] . A set of surface samples was collected at enrollment and every three days, including: 1) bathroom doorknob, 2) phone (all surfaces of the patient's phone and room phone), 3) overbed table and chair (pooled), 4) bed (bed rail and pillow) and light switch or pullcord in patient's bedspace (pooled), and 5) toilet and sink faucet handles (pooled) (Supplementary Figure 1) . Surface samples were collected by thoroughly wiping each surface type using the rough side of a dry 6 cm x 6 cm Swiffer cloth (Swiffer®, Procter & Gamble, Toronto, Canada). Nasopharyngeal swabs and Swiffer cloths were immediately placed into universal transport medium (UTM; Copan Diagnostics, Murrietta, CA).

During the study period, four bioaerosol samplers were used for sampling the first 45 patients enrolled that were not intubated. For each patient, one to two different bioaerosol samplers were used in each run. Using an air sampling pump (GilAir Plus Personal Air Sampling Pump, Sensidyne, St.

Petersburg, FA), air samples were obtained using the 1 μm pore size, 37 mm polytetrafluoroethylene (PTFE) membrane filters (SKC Inc, Eighty Four, PA), the 37 mm three-piece cassette with 0.8 μm polycarbonate (PC) filter (Zefon International, Ocala, FA), and 25 mm gelatin membrane filters (SKC M a n u s c r i p t 6 Inc, Eighty Four, PA). Prior to sampling, the pumps were calibrated to a flow rate of 3.5 L/min using the corresponding filter used for sampling that day (Gilibrator 3, Standard Flow Dry Cell Calibrator, Sensidyne, St. Petersburg, FA). In the patient rooms, samplers were placed at 1 m and 2 m from the head of patient at the approximate level of and anterior to mouth and nose, with samples collected over a 2 h period. All filters were placed in coolers at the end of the sampling period for transport and processed immediately. Air samples were also collected using the NIOSH two-stage cyclone bioaerosol sampler (National Institute for Occupational Safety and Health, Morgantown, WV). The NIOSH cyclone bioaerosol sampler is comprised of stages collecting larger particles (>4 μm) in the first stage into 15 mL conical tubes, smaller particles (1-4 μm) in the second stage into 1.5 mL conical tubes, and particles <1 μm onto a PTFE filter. The NIOSH cyclone samplers were assembled in the laboratory in a biosafety cabinet and calibrated to a flow rate of 3.5 L/min (BIOS DC-1 DryCal flow calibrator, SKC Inc, Eighty-Four, PA). In the patient rooms, the sampler was placed 1m from the patient as above and sampling occurred over a 2 h period.

All samples were processed at Sunnybrook Research Institute on the day of collection.

Nasopharyngeal swabs and environmental samples were vortexed for 20 s before aliquoting and storage at -80°C. PTFE, PC, and gelatin membrane filters were placed in 3 ml transport media before being vortexed for 20s, followed by aliquoting and storage at -80°C. For the NIOSH cyclone bioaerosol sampler, 1 mL of transport media was added to the first stage, 500 μL to the second stage, CoV-2, the 5' untranslated region (UTR) and the envelope (E) gene, with human RNaseP as an internal control [21] . The cycling conditions were: 1 cycle of denaturation at 60 °C for 10 min then 95 °C for 2 min followed by 44 amplification cycles of 95°C for 10 s and 60°C for 15 s. Rotor-Gene Q software (Qiagen, https://www.qiagen.com) was used to determine cycle thresholds (Ct) and samples with Cts <40 in both UTR and E genes were considered positive. Correlation analysis indicates almost perfect correlation between Ct values for the UTR and E gene (0.99). We therefore present Ct values for the UTR gene target within the text; the Ct value results for both gene targets are summarized in contamination. The following variables were investigated: age, sex, Charlson comorbidity index [22] , smoking history, Clinical Frailty Score [23] , presence or absence of symptoms from onset to 24 hours post admission (cough, fever, diarrhea, delirium/confusion), hypoxia at admission (defined as oxygen saturation < 92%), admission to intensive care unit (ICU) at time of sample collection, use of exogenous oxygen during stay, prone position, receiving steroids for treatment on day of sampling, room type (regular private room or negative pressure room), and the presence or absence of a PCRpositive NP swab on or after environmental sampling date (PCR-positive NP swabs were further categorized to Ct>30 and Ct≤30); the sampling date refers to the date the sample was taken. If use of exogenous oxygen during stay was significant, oxygen delivery methods (intubation, facemask/nasal prong, high flow) were included to investigate individual oxygen requirements. Since samples were taken serially from each patient over the course of this study, we included onset of illness to sample date as a fixed-effect control to account for temporal variability. See Supplementary Table 2 for further variable details. The outcome of interest was SARS-CoV-2 PCR-positive environmental surface samples.

A causal diagram was constructed to examine possible confounding and intervening relationships among exploratory variables relative to a SARS-CoV-2 PCR-positive environmental surface sample.

Mixed-effects logistic regression models with a random intercept for unique patient identification to account for clustering were constructed using stepwise backwards elimination. Variables that were A c c e p t e d M a n u s c r i p t 9 significant, potential confounders, part of a significant interaction term, or a control variable (i.e.

onset of illness to sample date) were included in the final model.

Pearson and deviance residuals were explored for outlying observations. Model fit was assessed by determining if the best linear unbiased predictors (BLUPs) met the assumptions of normality and homogeneity of variance.

There were 78 inpatients with COVID-19 who consented to participate. All were confirmed to have COVID-19 with a positive nasal, mid-turbinate, or NP swab tested in a licensed diagnostic laboratory in Toronto prior to enrollment; diagnostic samples used for initial COVID-19 confirmation were not included in the present study. Patients were de-identified and randomly assigned a number from 1-78. 

A total of 219 follow-up NP swabs were collected. The median number of NP swabs collected per patient was two (IQR 1-4). Overall, 172 (79%) NP swabs from 74 (95%) patients were positive for SARS-CoV-2 by PCR ( Table 2) The median time between onset of illness and surface sample date was ten days (IQR 6-12) for all environmental surface swabs, nine days (IQR 5-12; range 3-20 days) for PCR-positive surface swabs, and four days (range 4-5) for culture positive surface swabs (Figure 2) . 

In the final multivariable mixed-effects model, the following were found to be associated with the detection of SARS-CoV-2 RNA in environmental samples: hypoxia on admission, PCR-positive NP swab with Ct ≤ 30 on or after the environmental sampling date, higher Charlson comorbidity index score, and shorter time from onset of illness to environmental sample date ( Table 3) .

The intraclass correlation coefficient between observations at the patient level was 53% (95% CI: 34-70%). No outlying observations were identified. Graphical exploration of the BLUPs for patient ID appeared to meet the assumptions of homogeneity of variance and normality.

In this prospective cohort study in Ontario, Canada, SARS-CoV-2 RNA was detected from surfaces (25%) and air (2%) in the acute care setting. The genomic analyses of whole SARS-CoV-2 sequences in the present work confirmed patients were the source of viral contamination of their immediate surroundings in this setting. Although direct comparison of our results to other studies is limited due to heterogeneity in sampling, processing and detection methodologies, proportionally higher rates of recovery of viral RNA from surfaces compared to the air are broadly consistent with other studies investigating SARS-CoV-2 surface and air contamination [1, [5] [6] [7] [10] [11] [12] .

A limited number of studies to date have recovered viable SARS-CoV-2 virus from environmental samples [14] . We attempted to recover SARS-CoV-2 virus from 36 environmental surface samples, A c c e p t e d M a n u s c r i p t 13 PCR-positive air samples were collected from within 1 m of the patient in three cases. However, we were unable to culture viable virus from any of these air samples. To our knowledge only one study has isolated SARS-CoV-2 from air samples in this setting [25] . However, it is important to note that the authors concentrated their samples prior to cell culture, potentially optimizing viable virus recovery from samples despite low concentrations of SARS-CoV-2 in the sample. Additionally, although no CPE was observed, Santarpia and colleagues did observe increases in viral RNA in cell culture [9] ; western blot and transmission election microscopy also showed evidence of viral proteins and intact virions. The difficulty in culturing virus from air samples likely relates to a combination of low viral concentrations, dilution effects, the effects of sampling itself on viral cell membrane and surface protein integrity [3] .

In the multivariable analysis, hypoxia on admission, a PCR-positive NP swab with a Ct of ≤30 on or after the environmental sampling date, higher Charlson co-morbidity index score, and shorter time from symptom onset to environmental sampling were significantly associated with the detection of SARS-CoV-2 RNA in environmental surface samples (Table 3) . Although, to our knowledge, no other study has investigated putative patient factors associated with environmental contamination using multi-variable modelling, our findings are consistent with several observational studies that show that viral load peaks in the first week of illness in COVID-19 patients, with active viral replication in the upper respiratory tract in the first five days of illness [1, 10, 26, 27] . Additionally, both hypoxia and a high Charlson comorbidity index have previously been found to be associated with higher SARS-CoV-2 viral loads in the nasopharynx [28, 29] .

Our study has several limitations. First, although we had a large number of surface and air samples, the samples were recovered from only 78 patients, resulting in a relatively small effective sample size when accounting for clustering. While this still facilitated an exploratory analysis, this limited the A c c e p t e d M a n u s c r i p t 14 power of our multivariable analysis as indicated by the wide confidence intervals for some of the significant variables in our final model. The small effective sample size also prohibited us from investigating factors associated with viable virus in environmental surface samples, including time from symptom onset. Second, the present work focused only on acute care inpatients, excluded critically ill individuals, and had first samples obtained several days after onset of illness. Working in acute care allowed us ready access to patient areas for sampling and clinical data to garner a granular understanding of environmental contamination in hospital settings; however, the generalizability of our findings to other settings is limited, particularly where room ventilation is highly variable such as homes, schools, long-term care residences, other workplaces, and public spaces, or where patients may be pre-symptomatic or early in the course of their disease. Because of the pandemic, we were not able to access individual rooms to measure air exchanges. It is important to note that these data were collected prior to the emergence of the SARS-CoV-2 variants of concern (VOCs) in late 2020. Both the alpha and delta variants (B.1.617.2) are transmitted more efficiently [30] , but the modes of transmission and the effect of the difference in the variants on surface and air contamination are unknown. As such, the results from the present work represents a baseline that can be used to understand the transmission dynamics of VOCs. There was no apparent difference in the yield of SARS-CoV-2 with different air samplers and filters, but our study was underpowered to detect such differences. Finally, we did not use a standard curve for our RT-PCR analysis and could not calculate the virus concentration per volume of air. Therefore, we were not able to estimate a limit of detection for our aerosol samples.

The findings of this study provide insights into surface and air contamination with SARS-CoV-2 in hospitalized COVID-19 patients. We found that SARS-CoV-2 RNA was detected on a minority of surfaces in COVID-19 patients' rooms and rarely from air samples, and only early in the course of their hospitalization. These data suggest environmental sources are unlikely to pose a major exposure 

The funding body had no role the design of the study or in the collection, analysis, and interpretation of data nor in the process of writing the present manuscript. 4) , mild-to-moderately frail (5-6), and severely frail (7) (8) A c c e p t e d M a n u s c r i p t 24 A c c e p t e d M a n u s c r i p t 30 Figure 3 

Investigating SARS-CoV-2 surface and air contamination in an acute healthcare setting during the peak of the COVID-19 pandemic in London

Hospital-acquired SARS-CoV-2 infection: Lessons for public health

Detection of air and surface contamination by SARS-CoV-2 in hospital rooms of infected patients

Severe acute respiratory syndrome coronavirus 2 RNA contamination of inanimate surfaces and virus viability in a health care emergency unit

Aerosol and surface distribution of severe acute respiratory syndrome coronavirus 2 in hospital wards

Aerodynamic analysis of SARS-CoV-2 in two Wuhan hospitals

Air, surface environmental, and personal protective equipment contamination by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from a symptomatic patient

Aerosol and surface contamination of SARS-CoV-2 observed in quarantine and isolation care

The infectious nature of patient-generated SARS-CoV-2 aerosol

SARS-CoV-2 RNA detection of hospital isolation wards hygiene monitoring during the Coronavirus Disease 2019 outbreak in a Chinese hospital

Environmental contamination by SARS-CoV-2 in a designated hospital for coronavirus disease 2019

Environmental contamination of SARS-CoV-2 in healthcare premises

Assessment of air contamination by SARS-CoV-2 in hospital settings

Contamination of air and surfaces in workplaces with SARS-CoV-2 virus: a systematic review. medRvix

Rapid SARS-CoV-2 whole-genome sequencing and analysis for informed public health decision-making in the Netherlands

High-throughput SARS-CoV-2 and host genome sequencing from single nasopharyngeal swabs. medRxiv

Whole genome sequencing and phylogenetic classification of Tunisian SARS-CoV-2 strains from patients of the Military Hospital in Tunis

A comparison of whole genome sequencing of SARS-CoV-2 using amplicon-based sequencing, random hexamers, and bait capture

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies

How to obtain a nasopharyngeal swab specimen

Real-time PCR-based SARS-CoV-2 detection in Canadian laboratories

A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation

A global clinical measure of fitness and frailty in elderly people

Global initiative on sharing all influenza data -from vision to reality

Viable SARS-CoV-2 in the air of a hospital room with COVID-19 patients

Virological assessment of hospitalized patients with COVID-2019

Epidemiologic features and clinical course of patients infected with SARS-CoV-2 in Singapore

Nasopharyngeal viral load predicts hypoxemia and disease outcome in admitted COVID-19 patients

Association of SARS-CoV-2 genomic load with COVID-19 patient outcomes

Inflection in prevalence of SARS-CoV-2 infections missing the N501Y mutation as a marker of rapid Delta (B.1.617.2) lineage expansion in Ontario, Canada. medRvix

A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t