Secondary Logo

Journal Logo

Clinical Science Aspects

Aerosol Exposure During Surgical Tracheotomy in SARS-CoV-2 Positive Patients

Loth, Andreas G.; Guderian, Daniela B.; Haake, Birgit; Zacharowski, Kai; Stöver, Timo; Leinung, Martin

Author Information
doi: 10.1097/SHK.0000000000001655



Our knowledge of the currently globally active pandemic, SARS-CoV-2, is growing constantly. Even though the virus is detected in several different body tissues and secretions, it seems undisputed that the majority of infections are caused by droplet and smear infections since virus DNA can regularly be detected in the nose and throat as well as the lower respiratory tract (1–3). As with known human-pathogenic corona viruses SARS-CoV and MERS, transmission through aerosols containing the virus is postulated (4, 5). Iatrogenic diagnostic or therapeutic interventions in the upper and lower airways seem to be associated with an increased risk of infection for medical staff (6, 7). Even though no structured risk analyses are available for the current SARS-CoV-2 pandemic, present data indicate that healthcare workers have a higher risk for infection than average (8, 9). A study on the SARS epidemic from 2002 and 2003 showed that the risk of infection for surgical staff was significantly increased in the context of a tracheotomy with an odds ratio of 4.2 (1.5:11.5) (10). It is assumed for the SARS-CoV-2 pathogen that health care professionals in the head and neck area are exposed to a higher risk of being infected with an aerosol-transmitted pathogens (6, 7, 11).

Experimental data suggest that potentially viral aerosols are released especially during surgical or interventional manipulations. In cases of severe COVID-19 disease, the need for long-term ventilation may require dilatative or surgical tracheotomy. While experimental data on aerosol development are available for anesthesiologic interventions such as mask ventilation (12), intubation and bronchoscopy required for dilatative tracheotomy (13), comparable data on surgical tracheotomy are not available. Therefore, we studied the semiquantitative evaluation of aerosol exposure during surgical tracheotomy under the conditions of LAF in the operating room (OR). In addition, we reviewed countermeasures to improve the protection of medical staff against airborne transmission of infections.


Breathing simulator test setup

The experiments were carried out with a breathing simulator. A silicone tube with a diameter of 1.6 cm was used as tracheal equivalent. This corresponds to the average diameter of the trachea of an adult (1.5–2 cm) (14). To mimic the surgical situation of a tracheostomy, we opened an area of 1.3 cm2. Ventilation was performed manually via a ventilation bag with a maximum tidal volume of 800 mL (Ambu Spur II, Ballerup, Denmark). We chose this tidal volume because even though being relatively high it still matches a physiologic tidal volume especially when the patient is under stress—such as in a situation of a tracheotomie in local anesthesia. The latter allowed for better visualization since there was more fog to be detected. The lung was simulated with a test lung for ventilators (Dräger SelfTestLung, Dräger, Lübeck, Germany)

Visualization of the airflow

To visualize the airflow, a fog generator (Eurolite N10, Eurolite, Germany) was connected to the input valve of the resuscitator bag and the lumen of the bag was filled with artificial fog (Smoke-Fluid “B”, Eurolite, Germany).


The experiments were performed under low light conditions in the OR. The emerging artificial fog was illuminated from the side with a slit lamp (DLP projector, NEC, Model LT380) in order to achieve a better contrast against the darker background due to the Tyndall effect. The breathing simulator with the tracheotomy model was set up on the operating table and the surgeon was working from the side next to the model under real surgical conditions (see Fig. 1).

Fig. 1:
Experimental test setup. 1: Surgeon, 2: breathing simulator with emerging artificial mist as a breath aerosol model (red arrows), 3: slit lamp for contrasting the mist via the Tyndall effect, 4: full-HD camera for video recording for later evaluation.

Video documentation and post-processing

Video documentation was done with a full-HD camera with a recording speed of 25 frames/s. The video files were post-processed using proprietary software based on LabView (National Instruments, Austin, Texas, USA) (Fig. 2). Only the green component of the RGB signal was used, as the fog contrasts particularly well in this spectral range. Flickering of the image sensor of the camera was compensated by a sliding averaging of 7 frames each. Subsequently—comparable to the digital subtraction angiography principle—a blank value (without exhalation) was subtracted from each frame of the video stream. The resulting difference images enabled the evaluation of the temporal course and the direction and speed of propagation of the aerosol. After manual masking of the non-relevant parts of the image, the density of the fog was quantified by averaging grey values. (Supplemental Digital Content:

Fig. 2:
Image post-processing to evaluate the aerosol propagation: The real images are very dark and show a disturbing noise when viewed closely. This can be compensated by color plan extraction and sliding averaging of seven single images each. Subsequently, an initial image without artificial fog is subtracted, comparable to a subtraction angiography. This gives the fog a much better contrast. After a manual masking, the density of the fog can be quantified by averaging the gray levels.

Experimental conditions

The following six experimental conditions were examined using the tracheotomy model with five repetitions for each condition (see Fig. 3):

  • 1. closed airway + expiration
  • 2. closed airway + coughing
  • 3. open airway + expiration + with laminar air flow (LAF)
  • 4. open airway + expiration + without LAF
  • 5. open airway + coughing + with LAF
  • 6. open airway + coughing + without LAF
Fig. 3:
Schematic presentation of the six examined conditions: The respiratory aerosol is examined as a function of the open or closed airway, the functional state of laminar air flow (LAF) and compared to coughing and breathing.

For conditions 1 and 2 the closure of the surgically opened airways was achieved by inserting and blocking a ventilation tube. For simulated exhalation, the Ambu-bag was manually compressed at a constant speed over a period of 1 s to approx. 50% of its size, resulting in a tidal volume of approx. 400 mL. Coughing was simulated by the maximum rapid compression of the entire volume of the Ambu-bag. The room ventilation system in an OR setup is characterized by a low-turbulence, vertical air flow directed from the ceiling to the floor (“laminar air flow” [LAF]). Conditions 3 and 4 were carried out with LAF switched on and conditions 5 and 6 with LAF switched off.

Statistical analysis

The statistical evaluation of the aerosol's propagation velocity under different test conditions and the fog density in the operator's head area were evaluated using the non-parametric Mann-Whitney U test for unconnected samples with Microsoft Excel (Version Excel für Mac 16.16.20).


With the help of the chosen model, the six test situations described above were successfully simulated and examined. A video of the test conditions can be downloaded via the following URL: Supplemental Digital Content,

Test conditions 1 and 2: closed airway

In test conditions 1 and 2 (exhalation and coughing with closed airway) no aerosol was detected even during expiration or coughing. The cuff of the ventilation tube sufficiently sealed the lower airways and bridged the area of the tracheostomy.

Test conditions 3 and 4: expiration with open airway

The two upper image series in Figure 4 illustrate the distribution of the aerosol after exhalation with an open tracheostomy. In test condition 3, with activated laminar flow, the aerosol rises within 1.6 s up to about 40 cm above the tracheostomy. However, this is slowed down continuously by LAF and finally even reverses into a downward movement. This forces the main part of the aerosol cloud downward and away from the surgeon. Only very small amounts of the aerosol are exposed to the surgeon's face area.

Fig. 4:
Image series of test conditions with open airway: The upper limit of the aerosol cloud is marked with a white line from picture to picture to illustrate the direction and speed of propagation. Note the different time scales of the series for breathing and coughing given at the bottom left of each image. LAF indicates laminar air flow.

In test condition 4, the effects of a missing LAF are simulated: upward movement of the aerosol is only slowed down by gravity and thus considerably slower. There is no reversal of the aerosol movement and the facial area of the surgeon is exposed to considerable amounts of the aerosol.

Test conditions 5 and 6: coughing with open airway

The two lower image series in Figure 4 visualize the aerosol propagation after coughing. Within less than half a second the aerosol reaches the surgeon's face area. In this time window, the aerosol propagation is a linear movement. No influence of the LAF on the aerosol propagation could be demonstrated: The speed of propagation at the coughing condition was 88.6 ± 5.9 cm/s (mean values ± standard deviation) and 95.0 ± 6.8 cm/s (orange columns in Fig. 5, left). This was significantly higher compared to the expiration condition (P < 0.01). During exhalation, the aerosol spreads upward with 19.2 ± 7.5 cm/s or 24.0 ± 1.1 cm/s only (blue columns). Again, there was no significant influence of the function of LAF. However, the influence of the ventilation system is clearly shown in the comparison of the maximum particle density in the area of the surgeon's face area (right panel of Fig. 5). Without laminar flow, the facial area of the surgeon was exposed to 47.9 ± 10.8% of the aerosol during exhalation. With laminar flow, however, only 4.8 ± 3.4% was detected. This difference is highly significant (P < 0.01). Ventilation technology has no influence on the aerosol density during coughing (77.1 ± 9.8% vs. 76.0 ± 8.0%).

Fig. 5:
Representation of the propagation velocity (left image) and relative aerosol density in the mouth protection area of the surgeon (right image) for test conditions 3–6; statistical significance is indicated above the columns (∗∗: highly significant with P < 0.01; Ø: no significance).


Aerosols—as expected—are the main transmission causes for SARS-CoV-2 infection. This leads to the need of a closer examination of specialist areas, which are active in the upper respiratory area and thus are exposed to aerosols. While experimental data on aerosol exposition are available for anesthesiologic interventions such as mask ventilation, intubation and bronchoscopy, comparable data on surgical tracheotomy are not existing to our knowledge (12, 13, 15). We determined the aerosol distribution during conventional tracheotomy and developed potential measures to protect against aerosol exposure. For this purpose, a manual breath simulator filled with artificial fog was used in this study. The distribution of the aerosol was video-documented and highlighted by subsequent image processing so that even highly diluted artificial fog could be detected and analyzed.

Closed airway (test conditions 1 and 2)

As the most important result of the presented study we demonstrated that no detectable amounts of aerosol are released from an opened trachea when the tube was inserted and sufficiently blocked. This was evident both for regular breathing and for the cough condition. Sealing the trachea with an inflated balloon below the tracheostomy site minimized aerosol exposure to personnel involved. Hence, the best protection against airborne pathogen transmission during surgical tracheotomy is an intubation of the patient. However, intubation itself also carries the risk of infection and protective mesures have already been presented which should make the intubation process safer (16, 17). When intubation is not possible, coniotomy followed by surgical tracheotomy may be considered (18).

Preserving cuff integrity is one of the most important protective measures to reduce exposure to aerosols. If the cuff is damaged, re-intubation via the tracheostomy and finishing the sutures of the tracheostomy only after sealing is renewed may be considered.

Breathing with open airway with and without LAF (test conditions 3 and 4)

When the trachea is surgically opened, aerosol is detectable under respiration; during expiration, aerosols are almost vertically exposed toward the surgeon face area. Without a LAF, the upward movement of the aerosol is only slowed down by friction and gravity. Even if artificial fog is only a model for the respiratory aerosol, it can be assumed that the aerosol exposition presents a higher risk of infection.

The effect induced by the laminar flow was impressive as the propagation of the aerosol was successively slowed down by the outflowing air in the OR and the direction of aerosol movement is finally reversed. Although aerosol components still reach the surgeon through turbulences in the peripheral area, these are diluted by a factor of 10. As a result, tracheotomies should be performed in an operating room with proper air flow technology whenever possible, in order to reduce aerosol propagation and to avoid vortex formation. The ideal setting would be a negative pressure operating room with LAF as described for example by Chow et al. (19). However, this is not available in all hospitals treating SARS-CoV positive patients. Several authors recommend staying in the intensive care unit to avoid virus distribution during transport (18, 20–22).

In light of our results this recommendation must be reviewed as LAF is not regularly available in intensive care units. In this respect, the decision whether a patient should be transferred to the OR should be made on a case-by-case basis and according to local conditions.

Coughing with open airway with and without LAF (test conditions 5 and 6)

In contrast to normal breathing, a cough-related aerosol propagation was practically unstoppable. No relevant delay of the aerosol's propagation speed by the laminar flow was detected. Velocities of up to 3.6 km/h (2.25 miles/h) were detected. In the literature, cough induced air speed of up to 200 km/h (125 miles/h) and ranges of several meters are reported (15, 23). The driving force behind this is the pressure build-up in the lower airways when the glottis is closed, followed by a glottis opening and a sudden drop in pressure. This mechanism is no longer possible after opening of the trachea, which explains the 55 times slower ejection speed during tracheotomy. Interestingly, we have not been able to find any literature regarding the coughing speed in tracheostomized or laryngectomized patients. However, even this relatively slow cough speed is sufficient to transmit an aerosol-containing cloud in shortest time to the surgeon, potentially leading to a higher risk of infection. This conclusion is also shared by Tay et al. (20) in their analysis of tracheotomies during the SARS epidemic and the resulting recommendations for the current pandemic.

In clinical practice, this means that coughing should be avoided during tracheotomy by any means. This is especially important during re-intubating or changing the tracheal cannula. On the other hand, in case of a tracheotomy under local anesthesia, a sufficient analgosedation is desirable to suppress coughing as far as possible (20).

Limitations of the study

We used a simple lung simulator to minimize dead space and to optimize the videographic detection of the artificial fog. Therefore, the airflow during exhalation and coughing was created manually with the help of a ventilation bag. This instrument has been put to the test many times in daily anesthesiological routine to imitate natural physiological breathing. The selected tidal volumes and expiratory times, despite being relatively high, still correspond to the physiological values given in the literature (24). We choose high normal values on purpose to make up for possible non-compliant lungs in COVID-19 patients which will most likely lead to an increase of aerosol leakage through the tracheostomy site.

The used aerosol model (artificial fog) and the exhaled air of a person have a different composition. The artificial fog used is created by the evaporation of fog fluid consisting of a mixture of double-distilled water and polyethylene glycol. Simplified, this corresponds to steam that remains in the air longer due to the stabilizing glycol content. Steam has an average droplet size of 5 to 10 μm. Physiological breath aerosols can also contain smaller components whose sinking rate is slower and which can therefore be spread further (23). Even the breath aerosols during expiration and coughing are not identical: Thus, larger droplets are released during coughing, which could carry more viral material, but also sink faster due to the higher weight (25).

Despite these chemical differences, artificial fog is established as a model for exhaled air in scientific studies (12). The steam used here with a sinking speed of approx. 10 mm/s without external air movement, is considered by the authors suitable for identifying and semi-quantitatively evaluating aerosol exposure in tracheostoma surgery.

Furthermore, we unfortunately were not able to test in a negative pressure room since they were all in use for treatment of COVID-19 patients. We think despite that lack in our study we still provide valuable data since negative pressure rooms and especially negative pressure ORs are rare in many countries strongly impacted by the SARS-CoV-2 pandemic.


We have developed a model for the visualization of aerosol exposure during surgical tracheotomy and successfully examined various scenarios. With the help of this model we were able to show that sealing the lower airways with a sufficiently blocked cuff offers the best protection against aerosol distribution. As a consequence, in time of Covid-19 tracheotomies with airway intubation should be preferred over tracheotomies in local anesthesia, if possible. However, even under ideal conditions, aerosol leakage may occur in situations such as re-intubation or in the event of complications as cuff damage. We demonstrated that a conventional OR laminar air flow system has a significant effect on aerosol exposure of the OR personal. Therefore, tracheotomies preferably should be performed under presence of a laminar air flow system in order to reduce the infection risk of OR personal. This is of particular importance not only to provide individual protection, but also to maintain the functionality of medical care system.


1. Zou L, Ruan F, Huang M, Liang L, Huang H, Hong Z, Yu J, Kang M, Song Y, Xia J, et al. SARS-CoV-2 viral load in upper respiratory specimens of infected patients. N Engl J Med 382 (12):1177–1179, 2020.
2. Mackenzie JS, Smith DW. COVID-19: a novel zoonotic disease caused by a coronavirus from China: what we know and what we don’t. Microbiol Australia 41 (1):45–50, 2020.
3. Kim ES, Chin BS, Kang CK, Kim NJ, Kang YM, Choi JP, Oh DH, Kim JH, Koh B, Kim SE, et al. Clinical course and outcomes of patients with severe acute respiratory syndrome coronavirus 2 infection: a preliminary report of the first 28 patients from the Korean Cohort Study on COVID-19. J Korean Med Sci 35 (13):e142, 2020.
4. Anderson RM, Fraser C, Ghani AC, Donnelly CA, Riley S, Ferguson NM, Leung GM, Lam TH, Hedley AJ. Epidemiology, transmission dynamics and control of SARS: the 2002-2003 epidemic. Philos Trans R Soc Lond B Biol Sci 359 (1447):1091–1105, 2004.
5. Memish ZA, Perlman S, Kerkhove MDV, Zumla A. Middle East respiratory syndrome. Lancet 395 (10229):1063–1077, 2020.
6. Thomas-Rüddel D, Winning J, Dickmann P, Ouart D, Kortgen A, Janssens U, Bauer M. Coronavirus disease 2019 (COVID-19): update für Anästhesisten und Intensivmediziner März 2020. Der Anaesthesist 69 (4):225–235, 2020.
7. Judson SD, Munster VJ. Nosocomial transmission of emerging viruses via aerosol-generating medical procedures. Viruses 11 (10):940, 2019.
8. r. Available at:, Accessed on April 20, 2020.
9. T. Available at: Accessed on April 20, 2020.
10. Tran K, Cimon K, Severn M, Pessoa-Silva CL, Conly J. Aerosol-generating procedures and risk of transmission of acute respiratory infections: a systematic review. CADTH Technol Overv 3:e3101, 2013.
11. Vukkadala N, Qian ZJ, Holsinger FC, Patel ZM, Rosenthal E. COVID-19 and the otolaryngologist: preliminary evidence-based review. Laryngoscope 2020; Online ahead of print.
12. Chan MTV, Chow BK, Lo T, Ko FW, Ng SS, Gin T, Hui DS. Exhaled air dispersion during bag-mask ventilation and sputum suctioning—implications for infection control. Sci Rep 8 (1):198, 2018.
13. Thompson KA, Pappachan JV, Bennett AM, Mittal H, Macken S, Dove BK, Nguyen-Van-Tam JS, Copley VR, O’Brien S, Hoffman P, et al. EASE Study Consortium. Influenza aerosols in UK hospitals during the H1N1 pandemic—the risk of aerosol generation during medical procedures. PLoS One 8 (2):e56278, 2013.
14. Schünke M, Schulte E, Schumacher U. PROMETHEUS Kopf, Hals und Neuroanatomie Hrsg. 4. Auflage. Stuttgart: Thieme 2015.
15. Seto WH. Airborne transmission and precautions: facts and myths. J Hosp Infect 89 (4):225–228, 2018.
16. Canelli R, Connor CW, Gonzalez M, Nozari A, Ortega R. Barrier enclosure during endotracheal intubation. N Engl J Med 382 (20):1957–1958, 2020.
17. Au Yong PS, Chen X. Reducing droplet spread during airway manipulation: lessons from the COVID-19 pandemic in Singapore. Br J Anaesth 125 (1):e176–e178, 2020.
18. Wei WI, Tuen HH, Ng RWM, Lam LK. Safe tracheostomy for patients with severe acute respiratory syndrome. Laryngoscope 113 (10):1777–1779, 2003.
19. Chow TT, Kwan A, Lin Z, Bai W. Conversion of operating theatre from positive to negative pressure environment. J Hosp Infect 64 (4):371–378, 2006.
20. Tay JK, Khoo ML-C, Loh WS. Surgical considerations for tracheostomy during the COVID-19 pandemic. JAMA Otolaryngol Head Neck Surg 146 (6):517–518, 2020.
21. Skoog H, Withrow K, Jeyarajan H, Greene B, Batra H, Cox D, Pierce A, Grayson JW, Carroll WR. Tracheotomy in the SARS-CoV-2 pandemic. Head Neck 42 (7):1392–1396, 2020.
22. Choi SY, Shin J, Park W, Choi N, Kim JS, Choi CI, Ko JH, Chung CR, Son YI, Jeong HS. Safe surgical tracheostomy during the COVID-19 pandemic: a protocol based on experiences with Middle East Respiratory Syndrome and COVID-19 outbreaks in South Korea. Oral Oncol 109:104861, 2020.
23. Bourouiba L. Turbulent gas clouds and respiratory pathogen emissions. JAMA 323 (18):1837–1838, 2020.
24. Klinke R, Pape H, Kurtz A, Silerbernagl S. Physiologie Hrsg. 6. Auflage. Thieme 2010.
25. Johnson GR, Morawska L. The mechanism of breath aerosol formation. J Aerosol Med Pulmonary Drug Deliv 22 (3):229–237, 2009.

Aerosol; airborne infection; risk analysis; SARS-CoV-2; tracheotomy

Supplemental Digital Content

Copyright © 2020 by the Shock Society