Sterile Field Contamination by Elastomeric Respirators Versus Surgical Masks and N95s

Background: Elastomeric respirators are reusable and reliable protection from infectious aerosol particles such as SARS-CoV-2. There is a lack of safety data for use in sterile elds limiting application to operating room settings where high-risk aerosol generating procedures are performed. We hypothesized an equivalent reduction in sterile eld contamination would be achieved using an elastomeric respirator covered by a surgical mask as compared to a standard surgical mask or N95. Methods: Randomized controlled crossover experiment with repeat measurement comparing microbial and aerosol contamination of operating room surfaces for elastomeric respirators, elastomeric respirators covered by a surgical mask, N95, surgical mask, and no mask. 80 experiments were performed by participants with randomized order and balanced crossover to all masking groups (n=16 per masking group). Participants executed droplet and aerosol generating procedures while wearing: (Group 1) elastomeric respirator with mask, (Group 2) elastomeric respirator only, (Group 3) N95, and (Group 4) surgical mask. Positive control was established with the participant unmasked (Group 5). Contamination was measured by microbial growth on settling plates and optical particle counters (0.2+ and 2 um+ particles). Results: There was a reduction in microbial contamination at the sterile eld (p<0.001) for all masks (Groups 1-4) compared to unmasked (Group 5). The mean colony forming units (CFU) at the sterile eld was 0 CFUs for elastomeric respirator (+/- mask), N95, and surgical mask versus unmasked growing 1.875 CFUs. Compared to the unmasked control, the elastomeric respirator (+/- mask), N95, and surgical masks all resulted in a -0.75 difference in contamination (95% CI -0.91 to -0.48, p < 0.001). No signicant difference in contamination between the elastomeric respirator (+/- mask) and a surgical mask was detected. No signicant difference in particle counts (0.2 µm+ room sterile elds from microbial contamination to a similar level as surgical masks and surgical N95s. Additionally, there is no measurable difference in aerosol contamination of sterile elds between elastomeric respirators, surgical masks, and N95s. Our results demonstrate that elastomeric respirators can be considered as an option for healthcare worker respiratory protection in sterile eld environments and consideration should be given to expanding their use in healthcare settings. This is of signicance throughout healthcare and especially in areas where obtaining a reliable supply of N95s is dicult or cost prohibitive. We anticipate this paper to be a starting point for future research into the use of elastomeric respirators in diverse healthcare settings including clinical, procedural, and hospitals worldwide. one-way ANOVA and the two-sample t-test.


Introduction
Identifying options for healthcare worker (HCW) respiratory protection during aerosol generating procedures (AGPs) such as intubation, tracheotomy, and aero-digestive procedures on patients with COVID-19 from acute respiratory syndrome coronavirus 2 (SARS-CoV2) has become an area of signi cant focus. Powered air-purifying respirators/controlled air-purifying respirators (PAPRs/CAPRs) and single-use N95 masks are currently the primary form of respiratory PPE used in United States institutions for high-risk AGPs on COVID-19 and high-risk/untested patients. N95 respirators lter at least 95% of particles > 0.3 µm and are the minimum level of respiratory personal protective equipment (PPE) recommended by the Centers for Disease Control (CDC) for use in AGPs. 1 However, supply chain issues have resulted in shortages and di culty obtaining adequate quantities of N95s for HCW protection.
PAPRs/CAPRs are a reusable option for respiratory PPE but are di cult to use by surgeons/proceduralists that need to use microscopes, loupes, headlights, etc. A third option for respiratory protection is elastomeric half-mask respirators (EHMR) tted with lters ranging from N95 to P100 (certi ed to lter 99.97% of 0.3 µm particles). 2 EHMRs have the favorable features of reusability, durability, and potential higher-level protection compared to disposable N95 respirators. This has recently resulted in attention and adoption of EHMRs across some U.S. healthcare systems. 3 In contrast to N95s which lter both inhaled and exhaled air, EHMRs lter inhaled air but exhaled air is emitted un ltered through a one-way valve. The EHMR exhalation valve reduces humidity, heat, and carbon dioxide retention within the respirator, resulting in improved comfort and decreased physiologic demands on the wearer. [4][5][6] However, the lack of ltration at the exhalation valve has raised concerns for the use of EHMRs in sterile elds. As a result, the CDC does not recommend EHMRs when sterility needs to be maintained. 7 Nxo published studies on EHMRs have reliably investigated the risk of contamination of sterile elds, or acceptable mitigation strategies.
This study was designed as a randomized controlled crossover experiment to compare sterile eld contamination by EHMRs (with and without mitigation) versus the standard of care surgical masks and N95s in an operating room setting. The hypothesis was that EHMRs covered by a surgical mask would provide equivalent protection to the sterile eld as a surgical mask or N95. Secondary endpoints included: reduction in contamination at the anesthesia eld, reduction in contamination at the nursing eld, and difference in contamination between EHMR, EHMR covered with a surgical mask (EHMRwM), N95, and a surgical mask. Participants performed aerosol generating activities in an operating room setting for 10 minutes while wearing either EHMR, EHMRwM, N95, surgical mask or no mask; during this time contamination throughout the room by droplets and aerosols was measured ( Figure 1).

Operating room characteristics
The room baseline characteristics included temperature mean 19
Contamination occurred more frequently in the sterile eld than the anesthesia or nursing locations. Contamination data at the anesthesia and nursing eld are shown in Table 1. There was no statistically signi cant difference in the microbial contamination between masking groups at these individual locations. Combining all locations, unmasked showed EHMRwM and N95s showed 1 colonization per 16 experiments, EHMR showed 1 colonization per 8 experiments, and surgical masks showed no colonization. The differences were not statistically signi cant, and no colonization occurred at the sterile eld with any mask (Group 1-4).

Particle Results
The mean total particles size 0.2 mm+ and 2 mm+ for each masking group at the sterile eld, anesthesia eld, and nursing eld are shown in Table 2. There was no statistically signi cant difference in number of 0.2 mm+ and 2.0 mm+ particles produced between the EHMRwM, EHMR, N95, surgical mask, and the unmasked scenarios. This was true for all locations tested (sterile eld, anesthesia eld, and nursing eld) for 0.2 µm+ and 2.0 µM+ particles ( Figure 3).
Comparing Microbiology and Particle Results: Particle concentration was not associated with microbial contamination. Mean particle counts for tests with contamination (0.2 µm+ particles = 900 (std = 1500) count/feet 3 ; 2.0 µm+ particles = 59 (std 83) count/feet 3 ) and those without contamination (0.2 µm+ particles = 1000 (std 1800) count/feet 3 ; 2.0 µm+ particles = 53 (std 76) count/feet 3 ) were compared. Mean particle concentration did not differ substantially between runs based on microbial contamination. The particle concentration measurements might not serve well as a surrogate measure for risk of infection. Their signi cance related to viral contamination is unknown.

Discussion
During the COVID-19 pandemic, protecting HCW from infectious aerosols with appropriate respiratory PPE is imperative but has become challenging. A goal of zero nosocomial COVID-19 infections amongst HCW is an essential to maintain the workforce and its morale. EHMRs can provide reliable and potentially enhanced respiratory protection compared to the N95 respirator, especially for healthcare workers performing high-risk APGs on a recurrent or regular basis. EHMRs may be of greatest bene t in the operating room where long procedures with high-risk AGP may be commonly performed by anesthesiologists, otolaryngologists, thoracic surgeons, pulmonologists, gastroenterologists, etc.
This study is the rst to evaluate the safety of EHMRs in sterile eld settings. Our study did not nd the exhalation valve to cause statistically signi cant increased droplet or aerosol creation in comparison to a standard surgical mask or N95. Compared to unmasked baseline, there was signi cant reduction in contamination at the sterile eld when a participant was wearing an EHMRwM, EHMR, N95, or surgical mask. The use of any mask (N95, surgical or EHMRs ± surgical mask) resulted in elimination of detectable microbial contamination at the sterile eld. Signi cantly, there was no detectable difference in contamination of the sterile eld when comparing a standard surgical mask to the EHMR, irrespective of whether the EHMR was covered. This nding supports that EHMRs may be safely used in an operating room environment without risking microbial contamination of the sterile eld.
An unexpected nding was the signi cant reduction in contamination created by use of an EHMR alone, even without mitigation by a covering surgical mask. This may be related to the internal shape of the EHMR; the chin piece of the mask covers or partially occludes the exhalation valve area. As a result, the exhaled air stream must move around multiple bends within the mask prior to reaching the exhalation valve. Droplets may be collected within the EHMR by inertial impaction wherein droplets above a certain size are unable to make the bend with the airstream and impact the surface of the mask rather than escaping via the exhalation valve. [8][9] This inertial impaction of droplets within the mask would prevent them from reaching and exiting the mask through the exhale valve. Droplets ³1 micrometers may be stopped by this mechanism, including bacterial and viral droplet nuclei. 10 Inertial impaction of droplets within the mask could explain the ndings of reduced contamination at the sterile eld by the EHMR compared to an unmasked situation although additional modeling is needed.
The ndings of this study have signi cant rami cations for respiratory protection programs for health care workers. Currently, the CDC provides guidance for the use of EHMRs for protection from infectious droplets and aerosol, with the exception that EHMRs are not appropriate for use in sterile elds. This study provides evidence that EHMRs may be used in operating room settings without increased risk of sterile eld contamination. As EHMRs are reusable, can provide higher levels of protection, and are more cost effective compared to N95 masks, their potential use has signi cant rami cations for healthcare worker protection. 2 This is especially true in areas where disposable N95s or PAPRs/CAPRs are limited by supply or cost.
It is reasonable to cover the EHMRs exhalation valve with a surgical mask. Surgical masks protect HCWs from splashes and the patient from droplets and aerosols exhaled by the surgeon. The mitigation technique of covering an EHMR with a surgical mask showed no difference in sterile eld contamination compared to a standard surgical mask, supporting its safety. The lack of difference for microbial contamination between EHMRwM versus EHMR alone, is not surprising as prior systematic review of the literature failed to show association between surgical masks and reduction in surgical site infections. [11][12] Although surgical masks are cleared by the FDA with the intention of protecting the local environment from the wearer's exhaled breath, the certi cation testing is performed in-vitro and its correlation to live procedural scenarios is not established. Despite this, the continued use of surgical masks has been warranted as they provide protection to the surgeon from splashes, pose no danger to the patient, and may provide some as yet unmeasured bene t to the patient. Similarly, although our study found no measurable reduction in sterile eld contamination between the EHMRwM versus EHMR alone, the use of a surgical mask cover over the EHMR during sterile procedures is reasonable to provide additional protection to the patient and to prevent accidental splash/soiling of the respirator itself.
Participant "surgeons" in the study included an attending surgeon, a resident surgeon, a licensed surgical physician assistant, and a medical student. Therefore, the results can be extrapolated to members of a standard surgical team that are familiar with aseptic technique and use of personal protective equipment in accordance with national guidelines. At our home institution, the EHMRwMs have been used in "cleancontaminated" otolaryngology high-risk APGs. It is of signi cance that there has been no hospital acquired COVID-19 infections in the personnel or patients with this use, albeit in association with other precautions which enhance safety.
There was no signi cant difference for particle generation (0.2 and 2.0 µm+) between the EHMR groups, N95, and the surgical mask group. Particle sizes 0.2 µm+ and 2.0 µm+ were selected to represent a range of viral and bacterial droplet nuclei associated with prolonged suspension in air (2-12 hours) and thus unlikely to be detected by settling plates. The absence of difference between masking groups may relate to high ACH of the operating rooms where rapid air turnover creates prompt dilution of aerosol particles. The measured air handling within the operating rooms was higher (average 28.77 ACH; 95% CI 23.51 -34.02) than the 15-20 ACH recommended by national guidelines, but in line with national normal ranges of 15 -40 ACH for operating rooms. 13 -14 This high lambda results in approximately 0.5 air changes per minute with removal of 90% of airborne contaminants in 5 minutes. Prior data supports that increased ACH is associated with reduction in contamination of air. 15 Additionally, the testing conditions were 10 minutes of measurement with greater than 50% of that time being non-exertional breathing. This is representative of real-life scenarios and clinically applicable. However, aerosol particle size and distribution has been shown to vary with expiratory activities and the high percentage of time spent with non-exertional breathing may have resulted in normalization of aerosol data across the test groups. [16][17] Consistent with this, during high aerosol generating events we saw brief spikes of increased particle generation that rapidly returned to baseline. This, in combination with high ACH, may account for our aerosol ndings. Future testing that focusing only on aerosol generation should be considered.
There are multiple potential errors in this study's conclusions. The study was powered to detect a reduction in contamination at the sterile eld compared to the unmasked baseline. A study with hundreds of tests would be needed to identify potential small differences between the masking groups. Although this study was modeled after prior studies from the CDC with smaller samples, a larger data set would improve the strength of conclusions. 18 Given the similarity in particle data between all 5 masking groups, a much larger sample focusing only on aerosol generating activities should be performed to potentially detect aerosol generation differences. In this study, only one type of EHMR was tested with a forwardfacing exhalation valve. The forward-facing valve is considered highest risk, but studies looking at EHMRs with inferiorly directed valves are warranted. Additional potential error comes from variation between participants in their performance of experimental tasks, oral/nasal ora, oral hygiene, timing of last meal, etc.

Conclusions
Elastomeric respirators are feasiable alternatives to surgical N95s when performing high-risk AGPs in the operating room during the COVID-19 pandemic, conforming to the spirit of the CDC guidance on using respiratory protective equipment that provides the maximal level protection to the healthcare worker. This study did not nd increased risk of surgical eld contamination or aerosolization with an EHMR mask compared to a standard surgical mask or N95.

Study design:
This study was approved by the Mayo Clinic Institutional Review Board. The methodology of this experimental study was adapted from the CDC study on PAPR/CAPRs sterile eld contamination. 18 Experimental design, operating room setup, participant characteristics, testing battery, data collection, and statistical methods are outlined below.
The study was designed as a randomized controlled crossover trial with repeat measurements for a total sample size of 80. Four participants trialed each mask in randomly assigned order with balanced crossover for a total of 4 repeated measurements per mask. This resulted in 80 experiments distributed to Setting: A licensed operating room within the institution hospital was utilized. Baseline room conditions were recorded including temperature, relative humidity, and pressure. The air supply was HEPA ltered and airow was measured using carbon dioxide as a tracer gas. The lambda in each room meet national minimum air exchange rate of >15-20 air changes per hour (ACH). 13 The operating room setup included a surgical bed, a Mayo stand placed each of the 3 locations, 3 3 optical counter sensors (one placed on each of the 3 Mayo stands), and sheep blood agar settling plates ( Figure 1). The surgical table was placed 32 inches high. Mayo stands and specimen collection devices were placed at a height of 42 inches based on the average height of males/females in the US and considering surgical loupes with a 16-inch working distance. 19 Three Mayo stands were placed as sterile eld, anesthesia eld, and a nursing eld. The sterile eld Mayo was immediately in front of the surgeon centered on the OR table. The anesthesia eld Mayo stand was at the head of the bed centered 4.5 feet from the center of the sterile eld. The nursing eld Mayo stand at the foot of the bed had its center 9 feet from the center of the sterile eld. Directly in front of the surgeon, and at forty-ve degrees to the left and right, a printed passage was posted to indicate where the subject was to turn their head and read during a standard battery of aerosol generating movements.

Experimental Steps:
For each test, participants wore an impervious surgical gown, sterile gloves, and disposable hair covers.
All subjects were t-tested for EHMR with P100 lters and N95s per OSHA national standards. During each experiment, the participant performed 10 minutes of droplet/aerosol generating activities for each of the 5 mask test groups. The mask test order was randomized. At test initiation, participants walked to the surgical eld. Facing forward, they read a standard-length passage ("rainbow passage"), performed two forceful coughs, performed two deep exhales/inhales through the nose, and two deep exhales/inhales through the mouth for a total of two minutes. These maneuvers were then repeated with the head turned 45 degrees to the right (2 minutes), head turned 45 degrees to the left (2 minutes), and looking directly at the sterile eld. To complete the 10 minutes, the participant looked directly at the sterile eld while performing normal breathing. After completion, the participant backed away and waited 2 minutes to allow suspended particles to settle onto agar plates. Agar plates were then collected using sterile technique. Between each test, the participant changed their sterile attire and masks while mayo stands and particle counters were disinfected with Oxyvir® solution (Diversity, Inc. Fort Mill, South Carolina, USA).

Microbial Contamination Measurement:
Microbial contamination was measured using sheep blood agar settling plates in each eld during each test. This included 2 agar plates the sterile eld (placed side-by-side one centimeter apart), 1 plate at the anesthesia eld, and 1 plate at the nursing eld. The sheep blood agar plates collected for 12 minutes (10 minutes experiment with 2 minute settling time). This allowed thorough sampling of droplet particles as droplets size 5 -100µm would be expected to settle within 5.8 seconds to <12 minutes. 20 After each iteration, the agar plates were collected sterilely and transferred to the main hospital laboratory. The agar plates were incubated per our institutional microbiology laboratory standard at 5% CO2 at 37 degrees Celsius. Counts of colony-forming units (CFU) were measured at 7 days.
Aerosol Contamination Measurement: To measure contamination by aerosol particles, an optical counter sensor (Dylos DC11000 Air Quality Monitor, Dylos Corp., Riverside, California, USA) was placed on each eld with the sampling port facing the participant surgeon. The optical counter was previously validated for measurement particles 0.2 -2.0µm (0.2µm+) and 2.0 -10.0µm (2.0µm+). Measurement of aerosol particles 0.2µm+ (viral aerosols) and 2.0µm+ (bacterial aerosols) was performed continuously at all 3 locations. Total collection time was 10 minutes for particle data which was analyzed for total particle count during the 10 minute testing time span.
Statistical Analysis: Statistical power analyses based on preliminary data and the assumption that reduction of 14% points was signi cant (assuming p<0.05) as this translated into reduction of 1 colonization per 7 surgeries. Thus, 80 samples were necessary to provide a detectable difference of p <0.05. The sample was too small to model the correlation structure among repeated runs by the same simulator. Repeat observations were assumed to behave independently. The incidence of colonization was compared between mask types by using the Pearson chi-square test. The mean particle concentration was compared between runs with and without colonization by using the two-sample t-test. The mean particle concentration was compared between mask types by using one-way ANOVA and the two-sample t-test.

Declarations
Author Contributions: Brittany E. Howard MD: This author has substantially contributed to study design, data acquisition, data analysis, interpretation of data, drafting of the manuscript, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Ryan M. Thorwarth MD: This author has substantially contributed to data acquisition, drafting of the manuscript, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Karam Abi Karam BS: This author has substantially contributed to data acquisition, drafting of the manuscript, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Sam L. Snider BS: This author has substantially contributed to data acquisition, drafting of the manuscript, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Erica Forzani PhD: This author has substantially contributed to study design, data acquisition, data analysis, interpretation of data, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Joseph G. Hentz MS: This author has substantially contributed to study design, data analysis, interpretation of data, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Erin H. Graf PhD: This author has substantially contributed to data analysis, interpretation of data, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Bhavesh Patel MD: This author has substantially contributed to study design, data acquisition, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Kelly J. McKay EMT-P: This author has substantially contributed to study design, data acquisition, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.
Michael L. Hinni MD: This author has substantially contributed to study design, manuscript revision, and nal manuscript approval. This author also agrees to be accountable for all aspects of the work and has worked to ensure that any questions related to accuracy or integrity have been fully investigated and resolved.    Particle results at sterile eld. Results of total particle counts/feet3 for 0.2 µm+ aerosol particles at the sterile eld. Mean diamonds shown for ANOVA analysis with the top and bottom of each diamond representing the (1-alpha)x100 con dence interval for each group comparing results between masks (n=16). No signi cant difference in mean aerosol particle creation between masks. Abbreviations: EHMR with mask = elastomeric half mask respirator covered by surgical mask