key: cord-0969373-c8k466u1
authors: Avari, Hamed; Hiebert, Ryan J.; Ryzynski, Agnes A.; Levy, Ariela; Nardi, Julie; Kanji-Jaffer, Hasina; Kiiza, Peter; Pinto, Ruxandra; Plenderleith, Simon W.; Fowler, Robert A.; Mbareche, Hamza; Mubareka, Samira
title: Quantitative Assessment of Viral Dispersion Associated with Respiratory Support Devices in a Simulated Critical Care Environment
date: 2021-05-01
journal: Am J Respir Crit Care Med
DOI: 10.1164/rccm.202008-3070oc
sha: c620fdc0664687dfe938bd3f247d90262e5214bf
doc_id: 969373
cord_uid: c8k466u1

Rationale: Patients with severe coronavirus disease (COVID-19) require supplemental oxygen and ventilatory support. It is unclear whether some respiratory support devices may increase the dispersion of infectious bioaerosols and thereby place healthcare workers at increased risk of infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Objectives: To quantitatively compare viral dispersion from invasive and noninvasive respiratory support modalities. Methods: This study used a simulated ICU room with a breathing-patient simulator exhaling nebulized bacteriophages from the lower respiratory tract with various respiratory support modalities: invasive ventilation (through an endotracheal tube with an inflated cuff connected to a mechanical ventilator), helmet ventilation with a positive end-expiratory pressure (PEEP) valve, noninvasive bilevel positive-pressure ventilation, nonrebreather face masks, high-flow nasal oxygen (HFNO), and nasal prongs. Measurements and Main Results: Invasive ventilation and helmet ventilation with a PEEP valve were associated with the lowest bacteriophage concentrations in the air, and HFNO and nasal prongs were associated with the highest concentrations. At the intubating position, bacteriophage concentrations associated with HFNO (2.66 × 10(4) plaque-forming units [PFU]/L of air sampled), nasal prongs (1.60 × 10(4) PFU/L of air sampled), nonrebreather face masks (7.87 × 10(2) PFU/L of air sampled), and bilevel positive airway pressure (1.91 × 10(2) PFU/L of air sampled) were significantly higher than those associated with invasive ventilation (P < 0.05 for each). The difference between bacteriophage concentrations associated with helmet ventilation with a PEEP valve (4.29 × 10(–1) PFU/L of air sampled) and bacteriophage concentrations associated with invasive ventilation was not statistically significant. Conclusions: These findings highlight the potential differential risk of dispersing virus among respiratory support devices and the importance of appropriate infection prevention and control practices and personal protective equipment for healthcare workers when caring for patients with transmissible respiratory viral infections such as SARS-CoV-2.

The ongoing coronavirus disease pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has resulted in an unprecedented number of patients requiring hospitalization globally. Over one-fifth of people hospitalized with COVID-19 require supplemental oxygen, and up to 10-25% may require ICU admission and ventilatory assistance (1) (2) (3) (4) .

One of the major healthcare challenges of the COVID-19 pandemic is the safe delivery of respiratory support. The volume of patients needing advanced respiratory support has led to shortages of mechanical ventilators in some jurisdictions (5, 6) . High-flow nasal oxygen (HFNO) therapy and noninvasive ventilation (NIV) have been widely used for patients with COVID-19 and may reduce the need for invasive ventilation (2, 7) . However, bioaerosol dispersion in the environment, which may increase the risk of SARS-CoV-2 transmission to others, has been a concern when using HFNO and NIV (8, 9) . During the severe acute respiratory syndrome outbreak in 2003, nurses caring for patients receiving NIV may have been at higher risk of infection (10) . During the COVID-19 pandemic, the World Health Organization has recommended that HFNO and NIV should be used in selected patients suffering from hypoxemic respiratory failure with airborne precautions in place because of the uncertain potential to generate infectious aerosols (11) . HFNO has emerged as a common therapy in one-third to two-thirds of critically ill patients with COVID- 19 in Wuhan (China) and Lombardi (Italy) (2, (12) (13) (14) .

There is an important need to investigate the infectious potential of viral bioaerosols by different modes of oxygen delivery and ventilatory support (15) . A limited number of studies are available; these include models investigating nonviral particles, such as smoke, respiratory droplets, water, yeast, and bacteria, or bedside bacterial dispersion data from critically ill patients to investigate aerosol dispersion using noninvasive respiratory support systems (16) (17) (18) (19) (20) . This lack of evidence about aerosol generation (21, 22) underscores the importance of quantitatively investigating viral dispersion/infectivity as a function of various respiratory treatment modalities in a critical care context.

A simulated negative-pressure ICU room with a patient bed (Hillrom) was established in a 4.5-m (length) 3 3-m (width) 3 2.75-m (height) space at Sunnybrook Health Sciences Centre. The room-ventilation parameters were adjusted to achieve a 22.5-Pa pressure difference through supply and return airflows of 3.7 and 7.3 m 3 /min, respectively, with 12 exhaust air exchanges/h. All exfiltrated air was high-efficiency particulate air (HEPA) filtered before entering the heating, ventilation, and air-conditioning system.

The patient simulator was developed to mimic a spontaneously breathing patient with mild-to-moderate respiratory distress and to exhale a constant breath-to-breath viral load. The design consisted of a SimMan Classic (Laerdal Medical) recommissioned mannequin with a head attached to a 2-L neoprene reservoir bag via a 15-mm connector to serve as the tracheobronchial tree. The esophagus was clamped to prevent inadvertent system leaks. The mannequin head closely replicated the anatomy of a male human head, inclusive of anatomical structures of the trachea, larynx, oropharynx, mouth, and nose that may contribute to flow patterns. The patient simulator was positioned in a standard hospital bed (Hillrom) with the head of the bed in the flat position.

An LTV 900 transport ventilator (Pulmonetic Systems, Inc.) with a HEPA filter was used to simulate patient breathing patterns. The ventilator allowed for continuous volume delivery. To mimic a spontaneously breathing patient with mildto-moderate respiratory distress, the ventilator supplying air to the lung-reservoir bag was set to a respiratory rate of 29 breaths/min and an inspiratory-to-expiratory ratio of 1:2 (23 Puritan Bennett) to confirm the absence of a flow leak in the system. The average and maximum airflow at the patient simulator's mouth was measured to be 34 and 60 L/min, respectively.

The aerosolized DNA bacteriophage fX174 (HER 36) was used to model bioaerosol dispersion from the lower respiratory tract into the environment. To introduce the bacteriophage in the model's lower respiratory tract, an in-line Aerogen Solo vibrating mesh nebulizer (Aerogen Inc.) was attached to the distal end of the ventilator's outflow tubing immediately before the 2-L neoprene reservoir bag. Comparison of Aerogen nebulizers with an atomizer (model 9302; TSI Inc.) and a sixjet Collison nebulizer (BGI) showed a similar recovery yield for the bacteriophage fX174 (24) . The bacteriophage solutions consisted of amplified bacteriophage lysate at a concentration of 1:27 in phosphatebuffered saline (PBS). Before each experiment, 6 ml of the bacteriophage solution at a viral concentration of approximately 10 8 plaque-forming units (PFU)/ml was added to the nebulizer.

A schematic of the experimental setup is presented in Figure E1 in the online supplement.

Six different respiratory scenarios were examined. The first scenario examined invasive ventilation through an

Scientific Knowledge on the Subject: Patients with severe coronavirus disease (COVID- 19) require supplemental oxygen and ventilatory support. It is unclear whether some respiratory support devices may increase the dispersion of infectious bioaerosols and thereby place healthcare workers at increased risk of infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).

Comparing viral dispersion from invasive and noninvasive respiratory support modalities, our findings highlight the potential differential risk of dispersing virus among respiratory support devices and the importance of appropriate infection prevention and control practices and personal protective equipment for healthcare workers when caring for patients with transmissible respiratory viral infections such as SARS-CoV-2.

endotracheal tube with an inflated cuff connected to a mechanical ventilator (Puritan Bennett 980 ventilator; Medtronic) that used a VT of 250 ml, a respiratory rate of 16 breaths/min, and an N99 HEPA filter attached to the expiratory port of the ventilator (control). This scenario examined invasive mechanical ventilation once the closed-loop system was established and did not examine the act of intubation itself. The second scenario examined helmet ventilation with a positive endexpiratory pressure (PEEP) valve (Subsalve), an oxygen-air mixer (MaxVenturi), gas-outlet pressure of 8 cm In addition to these six scenarios, the use of a helmet operated with a mechanical ventilator (Puritan Bennett 980 ventilator; Medtronic) with a VT of 250 ml and a respiratory rate of 16 breaths/min was also examined, and the results are presented separately in Tables E1-E3. Six different locations were chosen to collect air samples. Locations 1-3 (10 cm above the level of the mouth, 40 cm vertically above the mouth, and 40 cm above and 30 cm behind the mouth, respectively) were chosen in close proximity to the patient mouth, where a healthcare worker may be located during aerosol-generating medical procedures such as endotracheal intubation. Locations 4 and 5 (100 cm above the ground and beside the bed and 150 cm above the ground and beside the bed) were representative of an assisting healthcare worker's hands and face. Finally, location 6 (100 cm above the ground and 275 cm from the mouth) was toward the foot of the bed. See Figures 1  and E2 for the schematic of the simulation facility and clinical relevance of airsampling locations.

Air samples were collected using 37-mm cassettes (SKC Inc.) with 1.0-mm polytetrafluoroethylene filters for a duration of 10 minutes per location. The cassettes were connected to a constant-flow air-sampling pump (Gilian Gilair-5; Sensidyne Industrial Health and Safety Instrumentation). Pumps were calibrated (Gilian Gilibrator-3 calibrator; Sensidyne Industrial Health and Safety Instrumentation) before each experiment at a flow rate of 3.5 L/min. Two positive controls were taken from the nebulizer before and after each experiment to quantify the bacteriophage concentration in the nebulizer for all experiments. Two negative controls were taken at location 1 and location 6 before each experiment to ensure that no bacteriophage residue was present within the patient simulator or ambient environment. Collection of negative controls was conducted by sampling the air at these two locations for 10 minutes while nebulizing PBS. Immediately after the 10-minute airsampling period concluded, 3 ml of trypticase soy broth (TSB) was added to each filter by removing the top cap of the polytetrafluoroethylene cassette and pipetting through the inlet. Then the top cap was placed back on and the cassette was placed on a vortex at maximum velocity for 2 minutes. The cassettes were rotated and flipped upside down during vortexing to ensure that all the viral particles that may have adhered to the sidewalls of the cassette during sampling were recovered.

Plaque assays were conducted to quantify the amount of viable bacteriophage recovered from each experiment (25) . The bacteriophage fX174 (HER 36) and Escherichia coli (HER 1036) were used to perform plaque assays. fX174 and its E. coli host were obtained from the Félix d'Hérelle Reference Center for Bacterial Viruses (Laval University). Trypticase soy agar was used for the culture media. fX174 was propagated in TSB with its E. coli host at 37 8 C under agitation. Once an optical density at 600 nm reached 0.1, the bacteriophage was added to the culture broth and the incubation was left overnight. The bacteriophage lysate was then centrifuged at 3,500 revolutions/min for 15 minutes to remove any bacterial cells or debris. The bacteriophage-containing supernatant was then filtered using a 0.45-mm pore-size syringe filter and stored at 280 8 C. Bacteriophage amplifications produced approximately 3.0 3 10 9 PFU/ml, as determined by plaque assay. Plaque assays were conducted in duplicate using TSB, trypticase soy agar (1.5%), and TSB soft agar (0.7%).

To measure the concentration of 0.3-, 1.0-, 2.5-, and 5.0-mm particles (measured as the particle count/L of sampled air), an optical particle counter (Aerotek 9303; TSI Inc.) was used. Particle concentrations were measured before starting each experiment (as a baseline measurement and to ensure no residual particles remained in the room from a previous experiment) as well as at the end of each 10-minute nebulization.

The bacteriophage concentration obtained from each air sampler was normalized by the average of the two positive control bacteriophage concentrations. For ease of comparison, the normalized value was multiplied by 10 8 before base-10 logarithmic transformations were performed. Raw data expressed in PFU/L of sampled air are presented in Table E1 .

For graphical presentation, bacteriophage concentrations were described using means and SDs after using logarithmic transformation. The Kruskal-Wallis test was used at each location to test whether any statistically significant difference exists among the six modes and whether there is an overall difference, and each mode was compared with invasive ventilation. No adjustment for multiple comparisons was made. SAS version 9.4 (SAS Institute Inc.) was used for all analyses. Comparisons were considered significant at P < 0.05.

A summary of the results is presented in Figure 2 . Bacteriophage concentrations and a summary of statistical analysis are presented in Tables E1-E3. The bacteriophage concentrations in the air varied substantially with respiratory support intervention and position (relative to the mouth) in the room. Across all locations, the lowest concentrations of the bacteriophage were achieved with closedcircuit invasive ventilation and noninvasive helmet ventilation with a PEEP valve. HFNO was associated with the highest bacteriophage concentration (with the exception of location 1). Generally, across the six locations, bacteriophage concentrations were highest at the mouth and above the mouth (location 1 and location 2) and lowest at the end of the bed (location 6).

One of the major considerations for the administration of suitable respiratory modalities for patients in ICUs is the risk associated with the generation and dispersion of infectious pathogens from patients' airways to healthcare workers and other patients. In this simulated critical care setting, a quantitative approach was taken to compare the viral dispersion associated with six respiratory support interventions.

The results revealed that closed-circuit invasive ventilation and helmet ventilation with a PEEP valve were associated with the lowest amount of dispersion of the infectious bacteriophage and that use of HFNO and nasal prongs was associated with the highest amount of dispersion. In general, bacteriophage concentrations were highest at the mouth and nose of the patient simulator and decreased with increasing distance. These results highlight the importance of appropriate infection prevention and control practices and personal protective equipment for healthcare workers performing aerosol-generating medical procedures for patients with COVID-19 (26) .

To date, only a few studies have addressed the potential exposure risk for healthcare workers performing aerosolgenerating medical procedures using detection of the virus in the air surrounding a patient. Intubation is commonly cited as an aerosol-generating procedure (27) . We demonstrated that after an endotracheal tube is placed in the airway, the cuff is inflated, the tube is connected to a closedcircuit system, and a filter is connected to the exhalation port of the ventilator, very little aerosolized bacteriophage material was dispersed throughout the room.

NIV using a face mask can reduce the need for invasive ventilation in selected patients with acute respiratory failure (acute respiratory distress syndrome). However, about half of patients with acute respiratory distress syndrome will require a transition from NIV to invasive ventilation because of intolerance to the device, worsening oxygenation or ventilation. An alternative strategy to face-mask NIV is to provide noninvasive support via a helmet (28, 29) . The findings from the present work suggest that viral concentrations in the air around a patient treated with a helmet and a PEEP valve were very low, in agreement with other published literature (30, 31) .

Only a few studies have previously considered the dispersion of viruses associated with HFNO in comparison with other oxygenation and ventilation support devices. This is of particular importance during the COVID-19 pandemic because of the frequency of HFNO use and the potential benefit in reducing the need for mechanical ventilation, and also in reducing the need for mechanical ventilators, which have been in short supply for healthcare facilities in areas of the world suffering the greatest number of infections (5, 6, 32) .

In this study, the use of HFNO led to higher concentrations of the bacteriophage at key locations where healthcare workers are in contact with patients but was not associated with substantially more dispersion of the bacteriophage than oxygen delivery by nasal prongs. Previous studies of oxygen-support devices using an exhaledsmoke model have demonstrated variable distances of dispersion with HFNO, nonrebreather, and Venturi masks (17, 33) , as have studies using water and yeast dispersion during HFNO use (18, 19) . However, smoke, yeast, and water models may not reflect the behavior of viral pathogens (34) .

The concentrations of the bacteriophage associated with the use of nasal prongs (at an O 2 flow rate of 4 L/min) at all six locations were found to be higher than those associated with use of the nonrebreather face mask (at an O 2 flow rate of 15 L/min). This interesting, yet significant, finding of the present study suggests that the flow rate may not be the only responsible parameter in the dispersion of infectious particles. This behavior is in agreement with the findings of other studies (30, 35) and highlights the importance of other factors (e.g., interface seal quality between the device and patient) in addition to airflow when considering infectious-particle dispersion in the patient environment.

Previous studies of exhaled breath and cough aerosols from patients with various respiratory infections have shown prominent similarities in aerosol size distributions, with a predominance of pathogens in small particles (i.e., ,5 mm [36] ). In a recent study, the median particle diameter generated from the respiratory tracts of 10 healthy individuals exposed to various modes of oxygen delivery, including nasal prongs, face masks, HFNO, and bilevel positive airway pressure, was reported to be within 1.01-1.53 mm (20) . Hence, a vibrating mesh nebulizer that generates aerosols with a small particle size was deemed to be clinically relevant for the generation of the viral aerosol in our model.

As with any simulation study, the present work has limitations. First, even though fX174 and SARS-CoV-2 are unrelated viruses with dissimilar structural biology that may lead them to behave differently during aerosolization, fX174 provides a model that mimics viruses like SARS-CoV-2 more closely than nonbiological tracers can. E. coli bacteriophages generally pose low risk to humans and require less biocontainment. Bacteriophages as models for human viruses have been extensively used in various workplace settings to understand the behavior of viruses under different experimental conditions (37) (38) (39) (40) , making them a feasible surrogate for SARS-CoV-2 in this study. Using a variety of nonpathogenic (to humans) bacteriophages with different features (e.g., enveloped and nonenveloped, RNA and DNA, single-and double-stranded) represents an additional line of investigation for future studies (36) . Until then, these findings provide information to clinicians in critical care areas on the relative potential risks of aerosol dispersion with common respiratory support procedures. It is also important to underscore that experimentally generated, nebulized bioaerosols do not directly reflect the number and size distribution or infectivity of bioaerosols generated by patients with COVID-19. Second, the model lacks fluids or mucus in the patient simulator's airway, which may change viral dispersion properties. Third, in this study, a negativepressure room was used to mimic the realistic isolation room that would be used to house patients suffering from a highly transmissible infectious disease under airborne precautions. Hence, the results may not be generalized to a standard hospital room. Fourth, the experiments were performed in a simulated environment and therefore cannot encompass all dynamic aspects encountered during clinical care of patients (e.g., ventilator-circuit disconnects during noninvasive or invasive ventilation), which might increase the risk of viral dispersion. Fifth, the reservoir in which the bacteriophage was instilled was below the level of the endotracheal cuff, simulating more lower-respiratory-tract viral replication and not viral replication in the upper airways. Sixth, a single-limb circuit with NIV was used, whereas some guidelines recommend a dual-limb circuit when treating patients with COVID-19. Seventh, the present results do not imply that there is limited viral dispersion in all patients with COVID-19 who are intubated. These experiments do, however, provide a quantitative assessment of the relative risk of a surrogate pathogen dispersion by directly comparing various common respiratory support devices under controlled conditions to help clinicians and infection prevention and control practitioners with risk assessments. Eighth, throughout the experiments, the same medical mannequin head was used in the patient simulator; thus, the potential impacts of varying facial morphologies were not studied.

the Northwell COVID-19 Research Consortium. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York city area

Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study

Critical care utilization for the COVID-19 outbreak in Lombardy, Italy: early experience and forecast during an emergency response

Clinical course and outcomes of 344 intensive care patients with COVID-19

Asian Critical Care Clinical Trials Group. Intensive care management of coronavirus disease 2019 (COVID-19): challenges and recommendations

Critical supply shortages -the need for ventilators and personal protective equipment during the COVID-19 pandemic

High-flow nasal cannula in critically III patients with severe COVID-19

Anaesthesia and COVID-19: infection control

Aerosol generating procedures and risk of transmission of acute respiratory infections to healthcare workers: a systematic review

Transmission of severe acute respiratory syndrome during intubation and mechanical ventilation

World Health Organization. Clinical management of COVID-19

Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China

Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan, China

China Medical Treatment Expert Group for COVID-19. Clinical characteristics of coronavirus disease 2019 in China

High-flow aerosol-dispersing versus aerosolgenerating procedures

Exhaled air dispersion during high-flow nasal cannula therapy versus CPAP via different masks

Assessment of the potential for pathogen dispersal during high-flow nasal therapy

Comparison of high-flow nasal cannula versus oxygen face mask for environmental bacterial contamination in critically ill pneumonia patients: a randomized controlled crossover trial

The impact of high-flow nasal cannula (HFNC) on coughing distance: implications on its use during the novel coronavirus disease outbreak

Aerosol generation from the respiratory tract with various modes of oxygen delivery

Prevalence and clinical presentation of health care workers with symptoms of coronavirus disease 2019 in 2 Dutch hospitals during an early phase of the pandemic

Coronavirus disease 2019 (COVID-2019) infection among health care workers and implications for prevention measures in a tertiary hospital in Wuhan, China

Physiologic effects of noninvasive ventilation during acute lung injury

Comparison of five bacteriophages as models for viral aerosol studies

Plaque assay protocols

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

Nosocomial transmission of emerging viruses via aerosol-generating medical procedures

Effect of noninvasive ventilation delivered by helmet vs face mask on the rate of endotracheal intubation in patients with acute respiratory distress syndrome: a randomized clinical trial

Severity of respiratory failure and outcome of patients needing a ventilatory support in the emergency department during Italian novel coronavirus SARS-CoV2 outbreak: preliminary data on the role of Helmet CPAP and non-invasive positive pressure ventilation

Protecting healthcare workers from SARS-CoV-2 infection: practical indications

Exhaled air dispersion during noninvasive ventilation via helmets and a total facemask

Projecting the demand for ventilators at the peak of the COVID-19 outbreak in the USA

Airflow and droplet spreading around oxygen masks: a simulation model for infection control research

Measurements of airborne influenza virus in aerosol particles from human coughs

Aerosol dispersion during various respiratory therapies: a risk assessment model of nosocomial infection to health care workers

Particle sizes of infectious aerosols: implications for infection control

Ozone efficacy for the control of airborne viruses: bacteriophage and norovirus models

Resistance of aerosolized bacterial viruses to Four germicidal products

Resistance of aerosolized bacterial viruses to relative humidity and temperature

Ozone treatment in a wind tunnel for the reduction of airborne viruses in swine buildings

The authors thank Marie-Ève Dubuis and Nathalie Turgeon for their help with bacteriophage nebulization and plaqueassay quantification. They also thank Roman Tymchal, Susan DeSousa, and Jose Cornejo for technical assistance during this study.

Invasive ventilation and helmet ventilation with a PEEP valve were associated with the lowest amount of dispersion of the infectious bacteriophage, and HFNO and nasal prongs were associated with the highest amount of dispersion, in a simulated critical care setting. These findings highlight the potential differential risk of the aerosolizing virus among respiratory support devices and the importance of appropriate infection prevention and control practices and personal protective equipment for healthcare workers when caring for patients with COVID-19. n Author disclosures are available with the text of this article at www.atsjournals.org.