Aerosol generating procedures

Aerosol generating procedures

In the first post of this two part series, I went in depth into the basic science of droplets and aerosols, how they spread, and how that might impact our infection control practices. This post will take a closer look at aerosol generating procedures.

Honestly, I am not that big of a fan of the term “aerosol generating procedures”. A procedure can be incredibly high risk and require different PPE without producing aerosols. On the other hand, aerosols can be produced without making a procedure high risk. (As is discussed in part one, even talking produces some aerosols.) What we really care about is the risk of transmission of disease and what PPE is required to protect us from that disease. Those are two separate questions, but unfortunately, I don’t think we have enough evidence to address them separately at this point, so for the sake of this post I will continue to use the generally accepted terminology of “aerosol generating procedures”.

Some general caveats

Before discussing the procedures themselves, there are some important caveats to consider:

  1. All of the available evidence is an incredibly low level, with a high risk of bias. I think it is worth reviewing, but we should avoid making any definitive conclusions. 
  2. Some of these procedures are almost impossible to separate retrospectively. Did the provider get infected while intubating or suctioning the patient? Was it the CPR or the use of a BVM? The data is simply not good enough to make such granular distinctions. 
  3. Severity of illness is an important confounder. Patients placed on noninvasive ventilation are likely to be sicker and therefore also likely to have much higher viral loads. Simply avoiding noninvasive ventilation will not lower their viral load and therefore may not reduce staff infection rates. Personally, I think this is the most important consideration. Placing a foley catheter was associated with a high risk of disease transmission in SARS. (Loeb 2004) This isn’t an aerosol generating procedure, but it places the clinician in close contact with a sick patient for an extended period of time.
  4. Much of the basic science research focuses on the distance that exhalations make it from the patient, using smoke or water vapour to visualize the breath. As was discussed in the first post, when sick patients are coughing or sneezing, these distances are increased tremendously. More importantly, the visual spread of droplets is irrelevant if they are small enough to evaporate and therefore spread infectious particles through an airborne route.
  5. Again, there is an important distinction between the physical production of aerosols and the risk of transmission of disease. Some procedures might produce more aerosols, but not put healthcare workers at higher risk. Conversely, procedures might not actually produce airborne aerosols, but could still represent very high risks of disease transmission to healthcare workers. It is not always easy, but I try to separate evidence for aerosolization from evidence of transmission below.
  6. Many studies of aerosol generating procedures don’t present their data as the total number of aerosols produced, but instead as a change from baseline. As is discussed in the main post on aerosols, the baseline production of aerosols can be very high. (Rule 2018; Simmonds 2010; Thompson 2013) Not seeing an increase in aerosols produced does not mean they aren’t present, and is not necessarily an argument for using less stringent PPE.
  7. Things aren’t as simple as declaring something “airborne” and adjusting your PPE. PPE is generally tested at rest, in idealized situations, which may not translate to real life practice. For example, doing chest compressions in PPE is incredible hard work, and can resulting in torn and loosened equipment, or sweating that interferes with the function of the equipment. (Shao 2020) 60-90% of people who pass an N95 fit test at rest will fail the test during active chest compressions. (Hwang 2020)
  8. Essentially all the studies that I could find on clinical transmission of disease are from SARS. I find it strange that no one has looked at the effects of aerosol generating procedures in other infections, so I worry I may be missing data.
  9. This document is a work in progress. You could do a full systematic review for each individual procedure (and maybe someone should). I don’t think there is a lot more information about clinical disease transmission, but there is almost certainly more information about the mechanics, chance of leak, and degree of spread for each of these procedures. I will add more if I find it. If you have other studies, please send them my way.


Evidence of increased risk of transmission to healthcare workers? YES

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

The systematic review by Tran (2012) reports a pooled odds ratio of 6.6 (95% CI 2.3 – 18.9) from cohort studies and 6.6 (95% CI 4.1-10.6) from case control studies. Fowler (2004) was included in that systematic review, and reported an increased chance of getting SARS, with a RR of 21 (95% CI 5-93) for nurses assisting with intubation and a RR of 13 (95% CI 3-59) for all healthcare staff in the intubation. They specifically note that all staff were wearing full PPE including an N95, although not everyone wore face or eye protection.

Basic science

The simple act of intubation should not produce aerosols. Neither placing a laryngoscope in the larynx, nor placing an endotracheal tube into the trachea result in the high flows of air or the changes in surface tension required to produce aerosols. If a patient is not adequately paralyzed, intubation can produce aerosols through coughing. (Judson 2019) This is an important consideration, because the coughing during intubation is often less vigorous than the coughing of a relatively well patient. Therefore, if we are going to consider intubation to be aerosol generating, we really need to consider any coughing patient to be aerosol generating. Aerosols could also be produced through other interventions in the peri-intubation period, such as the use of high flow oxygen, suction, or bag valve mask ventilation. The other possibility is that intubation simply represents a very high risk moment for contact or droplet transmission, due to the prolonged and close exposure to the patient’s mucous membranes.

  • One study found an increase in aerosols (defined as less than 7.3 µm) during intubation, but less detectable virus than baseline. (Thompson 2013) The study only included 5 patients.

Bottom line

Intubation should be treated as a very high risk procedure (like everyone is already doing). Here are some posts on how to do it as safely as possible:

Chest compressions

Evidence of increased risk of transmission to healthcare workers? PROBABLY

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

There is one case control study that reports a statistically significant increase in transmission to healthcare workers exposed to chest compressions (odds ratio 4.52 95% CI 1.08 – 18.8). However, they note that almost all of these patients were intubated, which is a strong confounder. (Liu 2009) Conversely, in 2 available cohort studies the odds ratio is only 1.4 (95% CI 0.2-11.2). The massive confidence interval means harm is certainly possible, but the numbers are just way too small to tell. (Tran 2012) Most important, there is evidence of significant heterogeneity, with one trial indicating an increased risk (OR 3.0) and the other indicating CPR could actually be protective (OR 0.4). Thus 2 of the 3 available studies suggest CPR increases risk. In the outlying study, with an odds ratio of 0.4, there were exactly 3 nurses who happened to do chest compressions and not contract SARS. (Loeb 2004) An n of 3 is clearly not enough to make definitive statements such as “chest compressions are not an aerosol generating procedure”, but unfortunately those exactly the conclusions we are seeing circulated based on this data.

Basic science

I have not been able to find any studies looking at the dispersion of aerosols or droplets during CPR. Pushing on the chest clear creates significant pressures and airflows, which are exactly the conditions required to produce aerosols, so I think we should assume they are produced until we see science saying otherwise.

Update: There is a simulation study (pre-print, not yet peer reviewed) showing that CPR does spread droplets and/or aerosols from the lungs, even in the presence of a mask over the patient’s mouth and nose, although the LMA with a viral filter looks promising: (Ott 2020)

Update 1a: This paper has now been officially published, and we have some more details. (Ott 2020) They used both a simulation mannequin and a cadaver. In both cases, they created the aerosols using a nebulizer and just watched their dispersion with chest compressions. Compressions clearly have the ability to disperse aerosols, and placing a mask over the patient’s face doesn’t seem to help. No aerosols were seen when an LMA with a viral filter was placed. Most importantly, this study doesn’t address whether chest compressions actually produce aerosols (but the basic science covered in the main aerosol post strongly suggests that they would.)

Update #2: We have run a number of simulations with a mannequin hooked up to a ventilator, and I am routinely seeing flow rates between 80 and 100 L/min on the ventilator. Expelling air at 80 L/min from the lungs seems high risk for aerosols, especially when compared to many of the other procedures we consider to be aerosol generating.

Flow rates during chest compressions (on a mannequin)


For some reason, this procedure has generated much more controversy than many of the others on the list, so I will include some of the official recommendations:

The ILCOR statement is: “We suggest that chest compressions and cardiopulmonary resuscitation have the potential to generate aerosols (weak recommendation, very low certainty evidence).”

The Australasian College for Emergency Medicine, the Resuscitation Council (UK), and the New Zealand Resuscitation Coucil all suggest airborne precautions during chest compressions.

The AHA states that “the administration of CPR involves performing numerous aerosol-generating procedures, including chest compressions, positive pressure ventilation, and establishment of an advanced airway.” Their first recommendation is “before entering the scene, all rescuers should don PPE to guard against contact with both airborne and droplet particles. Consult individual health or emergency medical services (EMS) system standards as PPE recommendations may vary considerably on the basis of current epidemiologic data and availability.”

Bottom line

Unless we are presented with new evidence, CPR should be considered a high risk procedure, with a high potential for aerosol generation.

Bag valve mask ventilation

Evidence of increased risk of transmission to healthcare workers? YES

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

There is only a single study and it suggests that BVM prior to intubation does increase risk of infectious transmission (OR 2.8 95% CI 1.3-6.4). (Tran 2012; Loeb 2004) However, manual ventilation alone was not associated with a significantly increased risk (OR 1.3 95% 0.5-3.2), although I have a hard time picturing the clinical scenarios in which BVM ventilation without intubation was occurring during SARS. (Tran 2012) It is almost impossible to distinguish the risk of BVM ventilation from the risk of intubation or other airway maneuvers, which are almost always happening at the same time. There is also a wide variation in skill levels when it comes to BVM ventilation, meaning results will probably vary. 

Basic science

  • Studies of smoke dispersion show leaks of up to 20-30 cm. That might sound reassuring, as no one bags a patient with their face less than 30 cm from the patient’s mouth. However, smoke dispersion is a very poor surrogate for what we really want to know. It does not tell us whether we would be seeing large droplet (local contamination) or aerosol (airborne) contamination. If aerosols are produced, they will remain airborne, and spread much farther than the 30 cm described in these studied. (Exactly how far depends on the ventilation of the room, among many other factors). There does seem to be a component of skill, as different providers have different degrees of dispersion, but the numbers are too small for any definitive conclusions.(Chan 2013;Chan 2018)

Bottom line

My guess is that BVM ventilation does increase risk, but if done well, with an excellent 2 hand grip, monitored with end tidal capnography, and a viral filter in place, that risk is likely small based on the numbers presented here. The risk associated with BVM ventilations are not as high as the risk from intubation, so the strategy of avoiding BVM just to continue with high risk intubation attempts probably doesn’t make sense. That being said, it is hard to say much from a single retrospective study.

Noninvasive ventilation

Evidence of increased risk of transmission to healthcare workers? YES

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

During SARS, noninvasive ventilation was associated with an increased risk of transmission to healthcare workers. There are two studies, and when combined they show a statistically significant increase in risk (OR 3.1 95% CI 1.4-6.8). Unfortunately, those clinical reports provide no details about the type of noninvasive ventilation being offered or the physical set-up of the ventilator, and those details might be very important.

Basic science

There isn’t just one “noninvasive ventilation machine”, and the differences between the various set-ups make interpreting and applying this data very difficult. There are 3 general locations from which aerosols could be generated, but much of this risk can probably be mitigated. Aerosols (and larger infectious droplets) could come from a leak around the mask, from a leak in the ventilator circuit, or from the exhalation port. Unfortunately, most of the studies don’t specify the source of aerosol generation.

Although it is a primary source of aerosols in many of the studies below, the exhalation port should be easy to deal with. Choose a mask that allows a viral filter to be placed over the exhalation port or use a system with a closed circuit (which returns exhalations to the machine). If you want some pictures of what this might look like, check out this PulmCrit post.

Leaks in the circuit are no different than leaks you might have with an intubated patient on a ventilator. Ensuring connections are tight and reinforcing with tape can help, but the risk is really eliminated if a viral filter is placed immediately between the mask and the ventilator circuit.

Therefore, the primary risk that needs to be considered is the risk of leak from around the mask. This will depend on both the patient and the mask. Some leak is probably inevitable, but it is unclear how high a risk this leak represents. For example, is the risk is any higher than the risk from aerosols produced by a coughing patient. (My semi-educated guess is that the risk from CPAP with a well fitting mask in a calm patient is probably lower than the risk from a coughing patient, but I also think we under-estimate the risk from a coughing patient, even without these aerosol generating procedures.)

Some studies:

  • There have been multiple simulation based studies that look at the distribution of droplets around a patient on noninvasive ventilation. Using exhaled smoke, Hui (2006, 2009, and 2014) demonstrated leaks around the noninvasive mask that sent exhalation plumes from 50 cm up to 100 cm from the patient. The vast majority of smoke measured in these studies came from the designated exhalation ports, and so could be contained with a viral filter. There was a small amount of leakage around the mask in the studies, but the amount and distance was much less. Again, I don’t care much about the distance that smoke travels. That might give you some information about your risk from large respiratory droplets, but if aerosols are being produced, they remain suspended in air, and will travel long distances.
  • One study did look specifically for aerosols, and so no increase. However, they only included 6 patients. (Li 2017)
  • In another study, noninvasive ventilation did result in more aerosols (defined as less than 7.3 µm), but the amount of virus detected was lower than baseline. Again, the n was only 11. (Thompson 2013) I cannot see in the paper what device they were using for noninvasive ventilation, but I have been told it was an open (unfiltered) system, so the results would not extrapolate to how we want to use these systems during COVID.
  • Another study reported no increase in the number of droplets (including those of aerosol size) with noninvasive ventilation, but the droplet counts at baseline were already in the 10s of millions per cubic meter, so a small increase might not have been noticed, and increasing the dispersion of that many droplets could be problematic. (Simmonds 2010)
  • There are many different CPAP masks, which complicates this data. One trial demonstrates essentially no leak from a Quattro Air face mask up to 20 cm H2O (although this is in a mannequin that isn’t moving), but significant air dispersion from both nasal CPAP masks tested. (Hui 2019) Note: The Quattro Air full face mask has a unique design that continuously vents air. Other CPAP masks have been shown to produce large air leaks.
  • The helmet NIPPV set-ups look more promising. With a good seal at the neck, air dispersion is negligible. When there is a leak, the dispersion reaches about 27 cm, which is similar to most low flow oxygen devices discussed here. (Hui 2015)

Bottom line

There is probably an increased risk of transmission when using noninvasive ventilation, but it appears to be somewhat smaller than the risk with intubation, which is often the only other alternative. (These risks cannot be considered in isolation.) However, the risk from noninvasive ventilation continues over a much longer period of time. How much of that risk can be mitigated with appropriate PPE and negative pressure rooms, as well as how that risk compares to our other options, remains unclear. The impacts of different mask designs is also unclear at this time.

Nebulizer treatment

Evidence of increased risk of transmission to healthcare workers? Probably

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

The meta-analysis includes 3 trials with wildly different results (OR 6.6, 1.2, and 0.1). (Tran 2012) The cited trial that is supposed to show that nebulizers are protective doesn’t present this data at all, so I am not sure where that OR of 0.1 comes from. (Wong 2004) In the trial with an odds ratio of 1.2, nebulizer treatment was essentially never used (2% in the group that didn’t develop SARS vs 0% in the group that did), so I am not sure that study allows for strong conclusions, although it has the biggest number of patients overall. (Raboud 2010) Numbers were similarly small in the study showing increased risk, with 3 out of 5 healthcare workers exposed to nebulizers developing SARS while 5 of 27 healthcare workers not exposed to nebulizers developed SARS. 

Basic science

By definition, a nebulizer creates aerosols. That is the point. However, the aerosols coming out of the nebulizer itself will be sterile. The question is whether these extra aerosols become contaminated and then act as a vector for transmission. I don’t have a good answer for that question.

  • Aerosols will clearly increase, because that is what a nebulizer is supposed to do. One study saw an increase of small droplets numbers in the 100s of millions with nebulizer therapy. (Simonds 2010)
  • What we really care about is whether the aerosols contain pathogens. One study, which is only available as an abstract, found that nebulization significantly increased the creation of small droplets (or aerosols), although none cultured any virus or bacteria. (Li 2017) Another study showed that there were more aerosols with nebulizer treatment, but that less virus was detected than baseline. (N only 3.) (Thompson 2013)
  • Compared to other types of oxygen mask therapy, nebulizers seem to spread droplets the furthest. Smoke or droplets can be seen up to a meter away with normal breathing. (Hui 2014) Therefore, nebulization might increase the dispersion of any infectious aerosols the patient is already producing.
  • What is being nebulized might matter a lot. The viscosity of the fluid in the lungs can make a big difference in both the number and size of aerosols being produced. For example, nebulized saline has been shown to decrease aerosols, and has been suggested as a possible infection control strategy. (Fiegel 2006) I wonder whether this partially explains the divergent results seen above.

Bottom line

The nebulizer increases the number of aerosols in the air, because that is exactly what the device does, but it isn’t clear how many of those aerosols carry virus, or whether there is an increased risk to healthcare workers. Seeing as we have an alternative in MDIs that seems safer, I think we should err on the side of caution, but the real answer is we have no idea what the risk is.

High flow humidified nasal oxygen (HFNC)

Evidence of increased risk of transmission to healthcare workers? Evidence doesn’t exist

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

As far as I can tell, we do not have any studies looking at transmission rates on HFNC. The technology wasn’t used during SARS, and the only studies identified by Tran (2012) assessing transmission rates were all related to SARS.

Basic science

  • Spread of macroscopic droplets was increased from an average of 2.48 meters to an average of 2.91 meters with high flow nasal oxygen in 5 healthy volunteers. The maximum cough distance was 4.5 meters when on HFNC. This study did not look at aerosols. (Loh 2020)
  • One study looked at room contamination in 20 patients with confirmed gram negative bacterial pneumonia, comparing HFNC to a standard oxygen mask. They didn’t find any of the gram negative bacteria. Total bacteria count was statistically higher at 1.5 meters with the high flow nasal cannulae (but many comparisons were made, and the clinical importance is questionable). The most important caveat is that bacterial pneumonia is likely to predominate in the lung parenchyma, while viral pneumonia will tend to have upper airway involvement, and so will probably be higher risk. (Leung 2019)
  • Some people are talking about a study that was done by the manufacturer of one of these devices and only published on the manufacturer’s website, which shows the dispersion of particles is increased by HFNC, but that applying a surgical mask over the HFNC eliminates the risk. I don’t think we can put a lot of stock into the advertising material of a company when making these types of high risk decisions.  What that data shows is a decrease is velocity and distance of air travel. This will significantly reduce the spread of droplets, but would have less of an effect on aerosols, which follow air currents and don’t required high velocities to travel long distances.
  • The distance that air travels may not be significantly increased when comparing HFNC to normal nasal prongs, possibly due to the tighter seal. (Hui 2019)
  • Because humidified air increases droplet size, it has been argued that HFNC might even decrease aerosol spread. (Hui 2019)


The Surviving Sepsis Guidelines say: “For acute hypoxemic respiratory failure despite conventional oxygen therapy, we suggest using HFNC over conventional oxygen therapy (weak recommendation, low quality of evidence”

The Australian and New Zealand Intensive Care Society says: “High flow nasal oxygen (HFNO) therapy (in ICU): HFNO is a recommended therapy for hypoxia associated with COVID-19 disease, as long as staff are wearing optimal airborne PPE.” “The risk of airborne transmission to staff is low with well fitted newer HFNO systems when optimal PPE and other infection control precautions are being used. Negative pressure rooms are preferable for patients receiving HFNO therapy.”

Bottom line

It’s hard to know. This technology creates exactly what is required for small aerosols (high air flows over areas of liquid or mucous), but there is really no good evidence. Much like noninvasive ventilation, I think that you have to proceed as if there is a significantly increased risk. However, assuming you have appropriate PPE and a negative pressure room, it is unlikely that the risk is as high as intubation, which is our highest risk procedure. Judgement will be required.

Oxygen masks, nonrebreathers, and nasal prongs

Evidence of increased risk of transmission to healthcare workers? YES

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

Manipulation of an oxygen mask was actually the highest risk factor in 1 study (OR 17 95% CI 1.8-165), but the numbers were very small. (Loeb 2004) In 2 studies combined the risk is not statistically significant (OR 4.6 95% CI 0.6-32.5), but the point estimate favoured increased risk in both trials. (Tran 2012) One study indicated that “high flow oxygen” was associated with a not statistically significant lower risk, but I can’t see high flow oxygen being defined anywhere in the paper. (Raboud 2010) There is a delicate balance between leaving someone on high flow oxygen and moving early to intubation (which clearly increases risk). No studies looked at the newer filtered nonrebreather oxygen masks.

Basic science

  • Visual droplets exhaled when wearing a venturi mask will reach up to 40 cm from the patient. Depending on the rate of oxygen delivery, breathing normally with nasal prongs will result in droplets 30-45 cm from the patient. (Hui 2014)
  • Images of airflow clearly show significant leak around the edges of the masks, which will limit the benefit of the newer masks we see with viral filters attached. (Hui 2007) Note that the video camera in this study didn’t measure beyond 40 cm. Coughing and higher flows sent contaminated air at least that far.
  • The higher the oxygen flow, the further the contaminated air disperses. (Hui 2007)
  • Using smoke to visualize exhalation, with normal oxygen flows (not flush rate), during normal breathing (not coughing or sneezing), exhalations reach between 10 and 30 cm from the patient with a nonrebreather mask. (Ip 2007; Hui 2014)
  • There is some evidence that filtered nonrebreathers limit droplet spread. In a study with a single healthy volunteer breathing through the mask, less vapour was seen in the HiOX (filtered) mask. (Somogyi 2004) However, this study obviously has some severe limitations, as visible water vapour is not really what we care about and leak around the mask is likely to be significantly higher in sick patients with respiratory distress.
  • One study did not see an increase in droplets after oxygen mask application, but the droplet counts at baseline were already in the 10s of millions per cubic meter, so a small increase might not have been noticed, and increasing the dispersion of that many droplets could be problematic. (Simmonds 2010)
  • One point that would be easy to overlook is that if you are using a Venturi mask, lower oxygen flow rates actually result in much higher total air flow rates. (Slessarev 2006) How that impacts aerosolization is unclear, but the dispersion distance is a little further with low ozygen flows rates than high in Hui 2014.
Air dispersion around a nonrebreather from Hui 2007
Venturi flow rates from Slessarev 2006

Bottom line

My guess is that a lot of the increased risk of transmission comes from the simple fact that people requiring higher levels of oxygen are sicker and likely have higher viral loads. That being said, these devices create the conditions required for aerosol creation (higher flows of oxygen), and there is evidence of transmission, so I think COVID patients requiring higher flow oxygen therapy should be treated with full precautions, much like those treated with HFNC or noninvasive ventilation.


Evidence of increased risk of transmission to healthcare workers? Maybe

Is there evidence that this procedure can be done safely with a reduced level of PPE? No

Clinical transmission

One study indicated a significant increase in transmission when suctioning was done prior to intubation, but with very large confidence intervals (OR 13.8 (95% CI 1.2-161.7). (Loeb 2004) The two studies combined indicate no statistical difference, with an odds ratio of 3.5 (95% CI 0.5-24.6). (Tran 2012) Suctioning after intubation does not appear to be a huge risk (OR 1.3 95% CI 0.5-3.4). (Tran 2012)

There is a hugely important confounder in this data. Which patients require suctioning during intubation? Those producing large amounts of infectious secretions. The truth is, we don’t usually suction unless we have to, so I don’t think there is anything to change here. Once patients are intubated, ensuring you have a safe inline suction clearly makes sense.

Basic science

  • One study showed no increase in aerosols detected with oral suctioning. (N only 6.) (Li 2017)
  • Another study showed more aerosols (defined as less than 7.3 µm) with suctioning but less virus than baseline. (Thompson 2013)

Bottom line

Suctioning gets a little complicated, because suctioning can actually decrease the amount of material available to aerosolize, especially after intubation, but the procedure itself could increase exposure. Not every study indicates exactly what they meant by suctioning. Clearly, if you need to suction during intubation, you should do so. Suctioning will be required post-intubation, but we should use enclosed systems if at all possible.


Evidence of increased risk of transmission to healthcare workers? Unclear

Is there evidence that this procedure can be done safely with a reduced level of PPE? NO

Clinical transmission

There is just way too little data. The systematic review gives us a pooled odds ratio of 1.9, but the 95% CI extends from 0.2 to 14.2, because the n is just way too small. (Tran 2012)

Basic science

  • There is one study, which demonstrated an increase in aerosols with bronchoscopy and more virus was detected than at baseline, but there were only 3 patients included. (Thompson 2013). 

Bottom line

I think bronchoscopy needs to be treated as a high risk procedure, but there is very limited evidence.

Other Procedures

There are way too many things that we can do to patients to go through each individual procedure. Overall, I think the increased risk from a lot of these procedures is probably a result of the increased time spent with a sick patient, rather than anything specific about the procedure. For example, inserting a urinary catheter was associated with an increased risk of contracting SARS (RR 5.00 95% CI 2.44-1.023), but we don’t think SARS is spread through the urine. (Loeb 2004) This is probably just a marker of the time spent in a room with a critically ill patient. Similarly, being present during an ECG (OR 3.52) and inserting a peripheral IV (RR 3.24) were associated with an increased rate of transmission of SARS. (Loeb 2004; Raboud 2010) Although we don’t consider these things “aerosol generating procedures”, they are a good reminder that you don’t need to be doing one of these procedures to be at high risk. Critically ill patients should be managed with the highest precautions. (Remember, patients make their own aerosols by talking, coughing, and sneezing.)

A few other studies

There is a Chinese language study, so I can only read the abstract, which combines intubation, tracheotomy, airway care, and cardiac resuscitation, and concludes all these aerosol generating procedures were associated with an increased risk of transmission of SARS to healthcare workers (OR 6.2 95% CI 2 – 18). (Ma 2004)

Rule (2018) had healthcare workers wear sensors during influenza season, and they did not detect an increase in viral exposure during aerosol generating procedures (viral copy was 29-321, or 20-223 viruses/m3, as compared to viral copy of 17-631 during routine patient care). However, that being said, they detected virus containing aerosols in 10 of 25 participants after an aerosol generating procedure. (I think this speaks more to the fact that there is an important baseline risk, rather than there being a lack of risk from the procedures.) The aerosol generating procedures in this study were 29 nebulizer treatments, 1 intubation, 4 airway suctions, and 5 nasopharyngeal aspirations. 

In a similar study, detectors were used to sample for influenza RNA at baseline and during aerosol generating procedures in 39 patients diagnosed with H1N1. Overall, more small droplets were produced during aerosol generating procedures, but less virus was detected in all procedures except bronchoscopy. Of note, aerosols and viral RNA were very frequently identified at baseline. There were numerous problems with this study. The sample size is very limited (5 intubations, 14 deep suctions, and 3 bronchoscopies). The baseline measurements were sometimes taken during chest physiotherapy, which in other documents has been considered aerosol generating. They discounted any droplets over 7.3 µm, but as is discussed in the main droplet post, larger droplets are probably still associated with airborne spread. (Thompson 2013)


Despite the fact that there is almost no science to base our decisions on, it is incredibly important that we understand the underlying concepts. I do not want any staff member at my hospital to become ill while caring for our patients, and we are required to do a large number of high risk procedures every shift.

The science is almost impossible to interpret. Although many of these individual procedures have been associated with transmission to healthcare workers, the fact that procedures occur on sick patients is an important confounder to consider. For example, in one study intubation was found to have an odds ratio of 2.79 for transmission of SARS, but an APACHE II score of greater than 19 had an odds ratio of 17.05 and a PaO2 to FiO2 ratio of less than 60 had an odds ratio of 8.65. (Raboud 2010) In other words, really sick patients were a stronger predictor of healthcare workers contracting SARS than the procedure itself.

That is an important insight that should shape the way we manage these patients. We have been incredibly focused on the procedures, but procedures occur over short periods of time, which inherently limits viral exposure. Routine assessments, despite creating less exposure to aerosols and droplets per minute, might actually be higher risk as a function of the total cumulative time spent with a high risk patient. This can partially be addressed by limiting time in the room as much as possible. (I have been really impressed by ICUs using longer tubing to allow pumps to be situated outside of patient rooms.) However, it also probably means maintaining a higher standard of PPE at all times when managing our sickest patients.   

These procedures can’t be simply divided into dichotomous high or low risk categories. There is a spectrum. As was discussed in the main post, aerosols are bad because they can remain airborne for very long periods of time. However, they are small and transmit far less virus than larger droplets. Even when airborne transmission is possible, most infections still occur as a result of prolonged close contact with patients. 

An intervention like CPAP might increase the number of aerosols produced, and require the use of N95 masks and negative pressure rooms. However, that risk may still be less than the wide spread droplets that the patient was producing while coughing, or the intimate close contact with those droplets that occurs while intubating.

Our information is incredibly limited. We don’t have any data that compares different procedures. Is it higher risk to keep someone on CPAP for 3 days or to risk an intubation? Should we place an LMA during a cardiac arrest, risking the small chance of a leak, or do we take the much bigger risk up front and intubate the patient? No one knows. No one should be making definitive statements, because there are no definitive answers. We will need to adjust based on our available equipment, the potential benefit to the patient, and our own local environments and expertise.

Other FOAMed

I recorded a podcast that covers this material with Dr. David Hao on the Depth of Anesthesia podcast.



Chan MT, Chow BK, Chu L, Hui DS. Mask ventilation and dispersion of exhaled air. Am J Respir Crit Care Med. 2013;187(7):e12–e14. doi:10.1164/rccm.201201-0137im PMID: 23540885

Chan MTV, Chow BK, Lo T, et al. Exhaled air dispersion during bag-mask ventilation and sputum suctioning – Implications for infection control. Sci Rep. 2018;8(1):198. Published 2018 Jan 9. doi:10.1038/s41598-017-18614-1 PMID: 29317750

Chen WQ, Ling WH, Lu CY, et al. Which preventive measures might protect health care workers from SARS?. BMC Public Health. 2009;9:81. Published 2009 Mar 13. doi:10.1186/1471-2458-9-81 PMID: 19284644

Fiegel J, Clarke R, Edwards DA. Airborne infectious disease and the suppression of pulmonary bioaerosols. Drug Discov Today. 2006;11(1-2):51–57. doi:10.1016/S1359-6446(05)03687-1 PMID: 16478691

Fowler RA, Guest CB, Lapinsky SE, et al. Transmission of severe acute respiratory syndrome during intubation and mechanical ventilation. Am J Respir Crit Care Med. 2004;169(11):1198–1202. doi:10.1164/rccm.200305-715OC PMID: 14990393

Hui DS, Hall SD, Chan MT, et al. Noninvasive positive-pressure ventilation: An experimental model to assess air and particle dispersion. Chest. 2006;130(3):730–740. doi:10.1378/chest.130.3.730 PMID: 16963670

Hui DS, Hall SD, Chan MT, et al. Exhaled air dispersion during oxygen delivery via a simple oxygen mask. Chest. 2007;132(2):540–546. doi:10.1378/chest.07-0636 PMID: 17573505

Hui DS, Chow BK, Ng SS, et al. Exhaled air dispersion distances during noninvasive ventilation via different Respironics face masks. Chest. 2009;136(4):998–1005. doi:10.1378/chest.09-0434 PMID: 19411297

Hui DS, Chan MT, Chow B. Aerosol dispersion during various respiratory therapies: a risk assessment model of nosocomial infection to health care workers. Hong Kong Med J. 2014;20 Suppl 4:9–13. PMID: 25224111 

Hui DS, Chow BK, Lo T, et al. Exhaled air dispersion during noninvasive ventilation via helmets and a total facemask. Chest. 2015;147(5):1336–1343. doi:10.1378/chest.14-1934 PMID: 25392954

Hui DS, Chow BK, Lo T, et al. Exhaled air dispersion during high-flow nasal cannula therapy versus CPAP via different masks. Eur Respir J. 2019;53(4):1802339. Published 2019 Apr 11. doi:10.1183/13993003.02339-2018 PMID: 30705129

Hwang SY, Yoon H, Yoon A, et al. N95 filtering facepiece respirators do not reliably afford respiratory protection during chest compression: A simulation study. Am J Emerg Med. 2020;38(1):12‐17. doi:10.1016/j.ajem.2019.03.041 PMID: 30955924

Ip M, Tang JW, Hui DS, et al. Airflow and droplet spreading around oxygen masks: a simulation model for infection control research. Am J Infect Control. 2007;35(10):684–689. doi:10.1016/j.ajic.2007.05.007 PMID: 18063134

Judson SD, Munster VJ. Nosocomial Transmission of Emerging Viruses via Aerosol-Generating Medical Procedures. Viruses. 2019;11(10):940. Published 2019 Oct 12. doi:10.3390/v11100940 PMID: 31614743

Leung CCH, Joynt GM, Gomersall CD, et al. Comparison of high-flow nasal cannula versus oxygen face mask for environmental bacterial contamination in critically ill pneumonia patients: a randomized controlled crossover trial. J Hosp Infect. 2019;101(1):84–87. doi:10.1016/j.jhin.2018.10.007 PMID: 30336170

Li J, Leavey A, Yang W, et al. Defining Aerosol Generating Procedures and Pathogen Transmission Risks in Healthcare Settings . 2017; 4(suppl_1):S34-S35. [article]

Liu W, Tang F, Fang L, et al. Risk factors for SARS infection among hospital healthcare workers in Beijing: a case control study . 2009; 14:52-59.

Loeb M, McGeer A, Henry B, et al. SARS among critical care nurses, Toronto. Emerg Infect Dis. 2004;10(2):251–255. doi:10.3201/eid1002.030838 PMID: 15030692

Loh NW, Tan Y, Taculod J, et al. The impact of high-flow nasal cannula (HFNC) on coughing distance: implications on its use during the novel coronavirus disease outbreak [published online ahead of print, 2020 Mar 18]. Can J Anaesth. 2020;1–2. doi:10.1007/s12630-020-01634-3 PMID: 32189218

Ma HJ, Wang HW, Fang LQ, et al. Zhonghua Liu Xing Bing Xue Za Zhi. 2004;25(9):741–744. PMID: 15555351

Ott M, Krohn A, Jaki C, Schilling T, and Heymer J. CPR and COVID-19: Aerosol-spread during chest compressions. 2020. Pre-print article available here.

Ott M, Milazzo A, Liebau S, et al. Exploration of strategies to reduce aerosol-spread during chest compressions: A simulation and cadaver model Resuscitation. 2020; [article]

Raboud J, Shigayeva A, McGeer A, et al. Risk factors for SARS transmission from patients requiring intubation: a multicentre investigation in Toronto, Canada. PLoS One. 2010;5(5):e10717. Published 2010 May 19. doi:10.1371/journal.pone.0010717 PMID: 20502660

Rule AM, Apau O, Ahrenholz SH, et al. Healthcare personnel exposure in an emergency department during influenza season PLoS ONE. 2018; 13(8):e0203223-. [article]

Shao F, Sun P, Tang Z. Cardiopulmonary resuscitation of inpatients with severe COVID-19 pneumonia: the Wuhan experience Resuscitation. 2020; [article]

Simonds AK, Hanak A, Chatwin M, et al. Evaluation of droplet dispersion during non-invasive ventilation, oxygen therapy, nebuliser treatment and chest physiotherapy in clinical practice: implications for management of pandemic influenza and other airborne infections. Health Technol Assess. 2010;14(46):131–172. doi:10.3310/hta14460-02 PMID: 20923611

Slessarev M and Fisher J. Oxygen Administration in the Emergency Department: Choosing the Appropriate Dosage and the Technology. 2006. Israeli Journal of Emergency Medicine. 6(1):10-21.

Somogyi R, Vesely AE, Azami T, et al. Dispersal of respiratory droplets with open vs closed oxygen delivery masks: implications for the transmission of severe acute respiratory syndrome. Chest. 2004;125(3):1155–1157. doi:10.1378/chest.125.3.1155 PMID: 15006983

Thompson KA, Pappachan JV, Bennett AM, et al. Influenza aerosols in UK hospitals during the H1N1 (2009) pandemic–the risk of aerosol generation during medical procedures. PLoS One. 2013;8(2):e56278. doi:10.1371/journal.pone.0056278 PMID: 23418548

Tran K, Cimon K, Severn M, Pessoa-Silva CL, Conly J. Aerosol generating procedures and risk of transmission of acute respiratory infections to healthcare workers: a systematic review. PLoS One. 2012;7(4):e35797. PMID: 22563403

Wong TW, Lee CK, Tam W, et al. Cluster of SARS among medical students exposed to single patient, Hong Kong. Emerg Infect Dis. 2004;10(2):269–276. doi:10.3201/eid1002.030452 PMID: 15030696

Cite this article as:
Morgenstern, J. Aerosol generating procedures, First10EM, April 6, 2020. Available at:

Leave a Reply

25 thoughts on “Aerosol generating procedures”

%d bloggers like this: