Journal Pre-proof Experimental study on the exposure level of surgical staff to SARS-CoV-2 in operating rooms with mixing ventilation under negative pressure

The purpose of this study was to reveal the exposure level of surgical staff to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from the patient’s nose and wound during operations on COVID-19 patients . The tracer gas N 2 O is used to simulate SARS-CoV-2 from the patient’s nose and wound. In this study, concentration levels of tracer gas were measured in the breathing zones of these surgical staff in the operating room under three pressure difference conditions: -5 pa -15 pa and -25 pa compared to the adjunction room. These influencing factors on exposure level are analyzed in terms of ventilation efficiency and the thermal plume distribution characteristics of the patient. The results show that the assistant surgeon faces 4 to 12 times higher levels of exposure to SARS-CoV-2 than other surgical staff. Increasing the pressure difference between the OR lab and adjunction room can reduce the level of exposure for the main surgeon and assistant surgeon. Turning on the cooling fan of the endoscope imager may result in a higher exposure level for the assistant surgeon. Surgical nurses outside of the surgical microenvironment are exposed to similar contaminant concentration levels in the breathing zone as in the exhaust . However, the ventilation efficiency is not constant near the surgical patient or in the rest of the room and will vary with a change in pressure difference. This may suggest that the air may not be fully mixed in the surgical microenvironment.


Introduction
Since the outbreak of the coronavirus disease 2019 (COVID- 19) pandemic, health care workers in hospitals have been at high risk of being infected by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1][2][3][4]. The operating room (OR) is a setting where both patients and surgical staff may stay for a long time, which increases the infection risk of surgical staff. The average duration of the majority of operation types is beyond 2 hours [5].
Therefore, ORs have also received great attention, and researchers have continuously proposed measures to prevent surgical staff from being infected, including the utilization of personal protective equipment (PPE) [6][7][8], aerosol boxes [9], and stricter infection control methods [8]. In addition, some hospitals shared their experience in the prevention of infection control in ORs [3,10].
Whether airborne transmission is one of the main modes of transmission for SARS-CoV-2 has been intensively debated since the beginning of the pandemic [11][12][13]. However, with more research conducted, airborne transmission has been widely recognized as one of the main modes of transmission for SARS-CoV-2. For instance, some studies have found viral ribonucleic acid(RNA) in air samples taken from rooms where COVID-19 patients have J o u r n a l P r e -p r o o f stayed [14][15][16]; other studies have found evidence of aerosol transmission by reviewing past outbreak events [17][18][19][20]. Ventilation is then considered to be an effective means of avoiding infection to prevent the spread of airborne transmission virus [21].
Based on the consensus that airborne transmission is one of the main modes of transmission of COVID-19, several measures around the operating room to prevent cross-infection are proposed. Many studies have suggested transforming ventilation in ORs from positive to negative pressure to treat COVID-19 patients [3,[22][23][24][25][26]. Compared with positive pressure ventilation, the air outside ORs flows into ORs by penetrating through the cracks of doors and windows of the negative pressure operating room (NPOR), which prevents SARS-CoV-2 from spreading out of the OR.
In contrast, the pressure difference of positive pressure ventilation has no significant effect because there is little air penetrating the room to disturb the airflow distribution in a positive pressure room. At present, there is very little experimental research on the exposure risk of airborne transmission of COVID-19 in NPOR, except in a few studies using simulation [27,28]. Different countries have different recommendations and requirements regarding the value of the pressure difference of NPORs. For example, the technical specification of surgical cleaning units in China clearly states that patients infected with airborne diseases should be operated on in a negative pressure operating room, but the specification does not indicate a specific pressure value [29]. British ventilation guidelines suggest that the pressure in a negative pressure operating room should be at least -5 Pa [30]. Both Canadian and CDC guidelines suggest a differential pressure of -2.5 Pa in a negative pressure operating room [31,32].
Equipping PPE has been recommended since the COVID-19 pandemic to prevent surgical staff from being infected in the OR by aerosol viruses released by patients [6,33]. PPE, J o u r n a l P r e -p r o o f including respirators, can effectively reduce the wearer's exposure to pollutants. In some cases, nurses, who usually stay far from the patient during operations, should also wear PPE during surgery. However, more than half of OR surgical staff reported a decrease in overall comfort with PPE, and more than 80% of respondents reported increased surgical fatigue. In addition, this combination of PPE can lead to increased respiratory work, reduced vision, reduced touch, and heat stress [34][35][36]. However, the air change rate in the OR is generally higher than 20 air changes per hour (ACH), and surgical staff away from the operating table may not be exposed to high concentrations of contaminants. This makes it difficult for surgical staff to choose protection equipment that may result in different safety and comfort levels. Therefore, revealing the contaminated exposure level of the breathing zone of surgical staff in the OR is an important indicator to help the medical staff determine whether PPE should be worn.
To date, there have been few studies on the exposure levels of surgical staff to SARS-CoV-2 in ORs. Some of these studies statistically analyze the existing infection cases of surgical staff. In addition to some investigation studies, some experimental and simulation studies were also carried out. Loth Andreas G et al. [37], conducted an experimental study of aerosol exposure levels in patients during tracheotomy and found that 4.8 ± 3.4% of aerosols were removed from surgeons in laminar ORs, compared with ten times as much in nonlaminar ORs. Alex Murr et al. [38] . used optical granulometry to measure aerosol concentrations during endoscopic nasal surgery. Aerosol concentrations at the surgeon's position were significantly increased with the republic bit, a miniature defibrillator. Ban C.H. Sui et al. [39] used a nebulizer to simulate aerosol exposure during intubation and compared aerosol concentrations in the operating room and isolation room, and the results showed that aerosol exposure levels in both rooms were similar. Marc Garbey et al. [40]      The OR lab is equipped with several real medical equipment for ORs, including an anesthesia machine ( Fig. 1(a)), an ultrasonic imager ( Fig. 1(b)), two endoscope imagers ( Fig. 1(d)), two surgical lamps, and an operating table. For some of the other medical devices that do not produce heat, we use manufactured models instead, for instance, two medical ceiling pendants, an instrument table, and some storage cabinets.
Six thermal manikins were used to mimic surgical staff in an OR, including one patient, two surgeons, two nurses, and an anesthesiologist. The specific locations of these manikins can be found in Fig. 1. The surface temperature of the anesthesiologist and the patient can be controlled by the temperature control device. The patient's head, arms, and body were set at 33℃, 30℃, and 31℃, respectively. The head, arms, body, and legs of the anesthesiologist were set at 33℃, 30℃, 31℃ and 30℃, respectively. The heat power of all other manikins is constant power whose value is set according to ASHRAE 55-2020 [42], which can be found in Table 1.  [43] found a strong linear relation (r 2 > 0.9) between the mean concentrations of microparticles sized 0.7 and 3.5 μm and the tracer gas nitrous oxide (N2O) in an indoor ventilated environment. The same study also showed that N2O, 0.7 μm, and 3.5 μm particles follow almost identical patterns with only a 2%-9% difference in the J o u r n a l P r e -p r o o f normalized concentrations. Noakes et al. [44] showed good agreement between the behavior of N2O tracer gas and 3-5 μm particles in a hospital isolation room. Furthermore, tracer gas has been extensively used in health care-specific studies to assess ventilation system efficiency in removing contaminants [45][46][47]. In this study, N2O tracer gas was used to simulate coronaviruses released from the breath and wound of a surgical patient. Studies have shown that aerosolized blood droplets may carry airborne viruses during laparoscopic surgery to release intra-abdominal pressure [48]. Surgical lights will be used at this time, so this study will use two surgical lights throughout all measurements. The sources were released from the nose and wound of the patient separately in each case The standards of most countries only set a minimum air change frequency, which is between 18-22 ACH [49][50][51][52]. The influence of ACH change on air distribution in the operating room has also been studied, so a constant air supply volume of 20 ACHs was used in all these cases. Most standards specify the temperature range, from 18℃-24℃ [49][50][51][52]. In this study, the air temperature was 22±1 ℃ in the OR lab during measurements. As there is little information on recommended values for negative pressurized ORs, three negative values were used in this study: -5 Pa, -15 Pa, and -25 Pa. The change in pressure difference in the OR lab was achieved by regulating the air extract rate. In total, nine cases were investigated with tracer gas under conditions with or without a ventilation system, different heights of surgical lamps, and different indoor heat intensities. The measurement conditions for each case can be found in Table 2. The air supply volume of all cases is 4000 m 3 /h.

Measurement procedure and instrument
The tracer gas N2O, which was placed outside of the OR lab, was continually released through a plastic tube (ϕ33 mm). The flow rate of N2O is controlled by a gas rotameter AirDistSys5000 anemometers were used to measure air temperature and airspeed at 5×7=35 points in the OR lab, as shown in Fig. 2(b). The lowest measurement points were 10 cm higher than those of the patient. These anemometers measure airspeed and temperature with an accuracy of ± 0.02 m/s (accuracy of ± 0.2 ℃, respectively. The measurement plane is J o u r n a l P r e -p r o o f shown in Fig. 2(b). Each measurement lasted for 3 minutes with a sampling rate of 0.5 Hz and the average values of 90 data were used. Before each experiment, the OR lab was ventilated for two hours in advance, the tracer gas was turned on after the indoor temperature was stable and the wall temperature remained unchanged. The sampling time of the Innova 1302 photoacoustic monitor was approximately 60 s/channel, and six channels were measured in sequence, giving 6 min between each measurement at the same location. The total sampling time for one case with two repeated measurements was between 300 and 350 min. The clean-up procedure was performed between two cases to keep the level of N2O rate below 0.5 ppm, which should be extracted from the measurement value.

Statistics analysis
Descriptive statistics were used for the analysis of the sampled N2O levels at six measurement points. Skewness and kurtosis statistics were used to test the assumption of normality for the N2O levels. Levene's test for equality of variances was used to test the assumption of homogeneity of the variance. Given that the statistical assumptions were met, a J o u r n a l P r e -p r o o f 1-way analysis of variance (ANOVA) assessed whether the mean N2O levels were significantly different between the same measurement points for different cases. If the statistical assumptions were violated, nonparametric Kruskal-Wallis tests were used as an alternative for ANOVA. Kruskal-Wallis tests the stochastic dominance between groups, i.e., whether any randomly measured concentration of N2O (ppm) from one case is higher or lower than any random concentration from another group. As both ANOVA and Kruskal-Wallis methods do not specify which specific cases were significantly different, significant differences between every two cases for one position were examined using Mann-Whitney U tests in a post hoc fashion. The significance was defined as p < 0.05. All analyses were conducted using SPSS version 27. (IBM, Armonk, NY).

Ventilation effectiveness
Air change efficiency and local air change index category indicators were used to interpret the measured results in this study [53,54]. The air change efficiency, , is defined as the ratio  it may be lower due to air recirculation and higher due to a better air distribution method.   [55]. . It depends on the position, the flow state. The distribution of air pollution is generally a two-parameter lognormal distribution [56]. This phenomenon is very J o u r n a l P r e -p r o o f common in contaminant detection, which is explained by the theory of continuous random dilution. According to theory, pollutants are still in the process of being diluted rather than completely mixed [57]. In this study, we observed this phenomenon in the breathing zones of surgical staff near the surgical microenvironment, including an anesthesiologist, assistant surgeon, and main surgeon, which suggests that tracer gas is still in the intermediate stage of dilution when moving to these areas rather than being completely mixed with air. This may indicate that surgical staff who work in proximity to a COVID-19-infected patient may be exposed to a higher coronavirus concentration than those who work further from the patient in ORs with mixing ventilation under negative pressure. Table 3 shows the results of the intergroup posttest, indicating whether there was a statistically significant difference between the two groups. Points 1 and 2, regardless of whether the pollutant is released from the nose or the wound, show similar characteristics.

Influence of pressure difference on exposure level
This is easy to understand since both sites are far away from the patient. When the pressure difference changes from -5 Pa to -15 Pa, there is a significant difference between the two groups of data, but when the pressure difference changes from -15 Pa to -25 Pa, there is no significant difference between the two groups of data. According to the actual measured exhaust air volume, we know that there is a small difference in the actual ventilation volume with different pressure differences. The ventilation volume differences between -5 Pa and -15 Pa and between -15 Pa and -25 Pa are 102 m 3 /h and 74 m 3 /h, respectively. This difference is due to the relationship between the pressure difference and ventilation volume determined by the following equation:   That is, the greater the local air index is, the smaller the concentration. This can explain the influencing factors of indoor concentration at each point, including the actual ventilation volume and air distribution. However, for points 5 and 6, this conclusion does not apply. We did not observe the relationship between the local air change index and concentration data.
When the differential pressure changes from -15 Pa to -25 Pa, the concentration data at point 5 differ significantly regardless of where the tracer gas is released. It can be seen from the J o u r n a l P r e -p r o o f thermal plume results that the change of pressure difference can significantly change the characteristics of the patient's thermal plume, which may be due to the influence of airflow infiltration into the gap of the room envelope, and this influence is more intense for point 5 than the change of ventilation rate and the change of the same fraction efficiency. When tracer gas was released from the wound, the concentration data of point 6 did not differ significantly under different pressure differences. This may be due to the Coanda effect on the chest of the assistant surgeon after the thermal plume from the patient's wound is diverted to the assistant surgeon. This effect will not be affected by external airflow.

Other factors affecting exposure levels
In this study, we observed that pollutant exposure levels of the assistant surgeon were much higher than those of other medical staff, despite the proximity of the assistant surgeon to the main surgeon. Previous studies of smoke from laparoscopic surgery have also pointed to high levels of assistant surgeon exposure to smoke from patients' wounds [58]. Field measurements by et al. at St. Olav Hospital also found that assistant surgeons were significantly more sensitive to particulate matter of various sizes during certain surgeries than other medical staff [59]. In addition, some hypotheses about the influence of the cooling fan on thermal plume formation are proposed based on the experimental results. Moreover, there is much surgical equipment in the operating room, so the airflow distribution will be significantly affected.
The formation of the thermal plume in this complex surgical microenvironment needs further research. Research on human thermal plumes in the surgical microenvironment may face several challenges: the influence of ventilation airflow [60], the modeling of thermal plumes [61], human body movement [62], and the effect of environmental temperature [63].
These studies also suggest that the human thermal plume is sensitive to environmental parameters. Therefore, the formation of the thermal plume in a more complex environment needs more in-depth research through computational fluid dynamic (CFD) studies or theoretical studies.
This study only analyzed the distribution tracer gas in the operating room by mixing ventilation with thermal manikins. However, the movement of surgical staff in the operating room is very normal, and it is not clear whether personnel movement will have a great impact on coronavirus distribution in ORs.
The results of this study provide the concentration of tracer gas in the surgical staff breathing zone in the operating room, and the corresponding concentration attenuation can be obtained.
Therefore, the infection risk of the surgical staff can be calculated by using an accurate risk assessment model and virus source intensity. Based on the risk of infection, recommendations J o u r n a l P r e -p r o o f can be made to the surgical staff on how to choose PPE equipment. However, it is difficult to calculate the precise risk of infection as there is no clear dose-response relationship for inhalation of the SARS-CoV-2. Therefore, risk assessment was not considered in this study.

Conclusion
In this paper, the distribution of airborne contamination released by COVID-19 patients in the OR was studied using tracer gas. The contamination sources were the noses and the wound area of the patients, and the concentrations in the breathing zone of surgical staff in the OR were measured. In addition, the velocity distribution and temperature distribution of the patient's thermal plume were measured, and the airflow interference was compared and analyzed. The following conclusion can be drawn.
• Contaminant exposure levels of the sterile nurse and the distribution nurse outside the surgical microenvironment are affected mainly by the ventilation airflow rate. The exposure levels of surgical nurses are sensitive to the ventilation rate. Due to the difference in the local air change index, there are slight differences in the measured concentration of contaminants between each measured nurse, but they are all around the average exhaust concentration.
• The main surgeon, assistant surgeon, and anesthesiologist in the surgical microenvironment had higher exposure levels to the contamination concentrations in their breathing zones, regardless of whether the source was the nose or a wound. The highest exposure concentration may occur for the assistant surgeon, which can be 12 times higher than the exposure level of other surgical staff outside the surgical microenvironment.
• The cooling fan of the endoscope image, which is located nearby, may have a great effect on the exposure level of the assistant surgeon. This may result in asymmetry in J o u r n a l P r e -p r o o f the patient's thermal plume, which in turn results in higher concentrations of contaminants in the assistant surgeon's breathing zone than in that of the main surgeon.
• Increasing the differential pressure resulted in lower concentration levels in the breathing zone of the main surgeon and the assistant surgeon. When the contaminant emission source was the nose, the exposure concentration of both the main surgeon and assistant surgeon decreased. When the contaminant emission source is the wound, the concentration level of the assistant surgeon will not be affected by the pressure difference because the thermal plume is biased toward the assistant surgeon.
• It is recommended that patients with COVID-19 should be operated in a negativepressure ventilated operating room with a large differential pressure during the COVID-19 epidemic to reduce the risk of infection on surgical staff.