Is COVID-19 spread by aerosols and aerosol-generating procedures?

This week's episode will be featured on Anesthesiology News. They've been a go-to resource for me to learn about the latest developments in our field. Check them out at anesthesiologynews.com. Welcome back to Depth of Anesthesia. My name is David Hao and I'm an anesthesia resident at the Massachusetts General Hospital It's my absolute pleasure to welcome Justin Mortensen to the podcast today.

Justin's a community emergency physician who works in Toronto, and he went to medical school at the University of Western Ontario and did a master's in health professional education at Western. He then went on to a residency in family medicine and a fellowship in emergency medicine at Queens. I actually found his website first 10 E. M. when I was researching aerosol generating procedures. And I'll just read you a little bit about what he does.

So he provides critical appraisals of important EM and critical care topics. And this portion straight from his website. He says He occasionally does deep dives into the literature, exposing myths or simply trying to explain what we do, why we do what we do. And I think that sounds a lot like what we try to do here on Depth of Anesthesia. Dr. Morenson, welcome to the show.

It's an absolute pleasure to join you, man. First, I want to thank you for being here because I know these are incredibly difficult times and to take some time off to go over the work you've done. I really appreciate that. Yeah, it's even though they're difficult times, I think it's really important to take some time to make sure we all really understand what we're doing and and why we're doing it. A lot of our anxiety right now stems from the fact that we're getting all sorts of conflicting advice, but without necessarily having that underlying science.

And so I'm hoping that science can actually be a way to allay some of everybody's anxiety. I really love that. So we'll start with a quick case today. Our case today is that of a 34-year-old man with obesity and hypertension who presents to the OR and he's positive for COVID-19. The resident who is involved in the case is concerned about transmission of the disease and recommends avoidance of any aerosol generating procedures if possible.

Uh Justin, what do you think are some of the inherent claims to this case? Yeah, so I guess uh there's a a couple obvious ones. Uh number one is that our procedures do in fact create aerosols. Uh number two is that those aerosols can spread uh the disease or that it will increase the risk of transmission. And I guess inherent in that term, aerosol generating generating procedure is another claim, which is that The aerosol is the important part of transmission.

Whereas what we really care about is we get sick, not necessarily the air the aerosol, so we care most about transmission. So off the top of my head, I think those are the the biggest uh claims or things that we need to think about. To our listeners, what's the level of your agreement with the claims that Justin laid out? And what's the level of evidence for what you believe? So I think to start off, it would just be great to review what exactly an aerosol is because I think this, like you mentioned in the articles, a complicated question to answer, but I think it lays the foundation of of our of our talk today.

Yeah when I started I actually thought this would be the easiest thing. I did not expect to like to walk into a minefield. There there's an academic war out there about exactly what an aerosol is. But at the very base it it's pr it's pretty simple. It's any kind of liquid or solid that floats in the air.

So you can think dust or pollen or you can think fog or mist. It's something that is suspended in the air. Um and The reason that matters is if something stays suspended in the air, it can travel a very far distance. Um and obviously if it's infectious, that that's a big problem for us. Um so speaking in broad terms Aerosols are going to be much smaller particles because obviously you have to have the buoyant force override the gravity.

Whereas droplets, which is what we normally talk about, are much larger uh particles and As an end result, what happens is gravity overcomes any buoyanc and they fall to the ground near the patient. Honestly, I read probably a hundred different papers on this, and the cutoff between the two is entire entirely unclear. So anybody who says you know there's a strict cutoff at 10 micrometers or five micrometers uh probably just hasn't read the next paper down the line. But in broad terms, smaller means it's more likely to be spread by airborne Larger means that the droplets are gonna fall all over the patient and you're gonna get it by touching the patient and rubbing your eye.

Or if you're like me, you examine a kid and they cough right in your mouth and that's how you end up getting getting sick. That's that's probably the big uh distinction we need to know. And I think inherent to what you just mentioned is the fact that because the cutoff is such a point of contention A lot of these papers are completely arbitrary in the sense that if you use a cutoff of two versus five or ten or twenty, your results are completely different. Yeah. It surprisingly, there's not actually that much empiric evidence.

People don't go around measuring aerosols that much. Almost all of these papers are mathematical models, and it matters so much in your mathematical model exactly what number you punch in. So a lot of the things we hear, like these rules about you know, if you're two meters away from the patient, you're perfectly safe, are based on mathematical models that have big assumptions in them. Uh and that's important to know that those assumptions might be faulty. So one of the terms I've heard thrown around almost as much as supply chain is aerosol generating procedures What exactly is an aerosol generating procedure and what does that matter for us?

Yeah, I'll tell you, I actually sort of hate this term uh because what we really care about is transmission. And The truth is you can get sick without somebody actually producing aerosols. So I'll give you an example. Intubation always gets put on this list. But if somebody is paralyzed, just placing a a laryngoscope into a patient's mouth doesn't really physically have any possibility of producing aerosols.

Aerosols need airflow over top of a a a li liquid surface. Um so we Intubation really shouldn't produce aerosols, but it is incredibly high risk for transmission. And on the other hand, just talking actually produces aerosols, but almost nobody gets sick by standing three meters away because somebody's talking. So I sorta hate the term. But in general, we're using it to mean two different things.

And that's why I hate it. Number one, we're mean using it to mean a procedure that could take a patient who's normally just producing droplets, who's not putting you at risk when you're three or four or five meters away and transforming those droplets into smaller particles so that now all of a sudden the disease is airborne. And so that would be like putting on a high flow and so it breaks up the particles and makes them much much smaller. And on the other side, we're using the term also to describe situations which are very high risk for transmission. And I think intubation is the example example uh there.

So that's a really long-winded way of saying basically we're using it to describe situations where we think we might get sick. Perfect. And I know that so you mentioned aerosols and normal activities. What kind of evidence do we have that aerosols are generated on a daily basis just by doing the regular things that we do? Yeah, th there's a number of good studies out there that um put these little sensors around patients to measure aerosols.

And it it's really funny actually, if you go back historically, people said that you never created these smaller particles and then you read into the fine print and it turns out that the monitors back before nineteen eighty or nineteen ninety or so couldn't actually detect anything smaller than a hundred microns. And so saying that there was nothing there smaller than that size was sort of a ridiculous statement. But there's a few tables that I included in my um in my blog post. Um, but there's one I'm looking at uh right right now by the first author name is I don't know if I can even say it, do gwid du g U I D, but basically in a cough, you you see somewhere in the range of a thousand particles per every cough in the 10 micron range, and then it's sort of a bell-shaped curve around that.

In a sneeze, there's something like 350,000 particles in that range. And again, a bell-shaped curve that goes around there. So I they've measured it, and there is pretty good evidence that every time you cough or sneeze, there's a a fair number of these articles If you go back to just talking or breathing, they can measure these very small uh aerosols. But we're talking now, instead of 300,000, we're talking about uh maybe a handful, ten or a few hundred. Um and so that that's why people get are much more likely to get sick with a cough or a sneeze than they are just by breathing.

Interestingly, when they do these studies, these are just averages. And the more I I've personally dug into the literature, the more I hear about superseding events or super producers. Does that mean the amount of bioaerosols generated can differ from person to person? Yeah, this is a fascinating topic, and I don't know how to apply it medically yet, but there's one paper by uh Fiegel in 2006 where they just took um volunteers uh and monitored them over the period of uh six hours doing their normal kind of activities. Uh and some of them had as few as fourteen individual particles detected.

And and the highest one was 3,230. So the the it's it's a factor of almost a thousand uh or more between the lowest producers and the highest producers of aerosols. Um and I guess we probably know that there are spitters when he you get a close talk to somebody. Like there are people who produce just more mucus, more e everything. There's probably a reason for it.

But when you hear hear about these single individuals who are spreading COVID far and wide, I think this is a really important phenomenon that I don't know that we understand well enough to do anything about at this point. And what about vomiting? Because um you mentioned here that there was a vomiting SARS patient associated with nosocomial spread in a hospital in Hong Kong. And I know that Up on the OB floor, there's concern that some of these patients who are in labor and vomiting might be generating aerosols. Yeah, it had a hard time putting together with vomiting.

Uh it makes sense, so you have high pressure and you have movement of fluids, so it physically should produce aerosols. Uh the best evidence was probably at just one bad case report from SARS. And it did appear that the uh the way infection spread through the unit looked more like it patter followed the ventilatory patterns rather than just what you would expect if it was from touching it would be a random pattern around the floor. But that's not very strong evidence, uh, because it could easily be that just one nurse walked in that direction, touched something and s and spread it along the way. So uh it's likely that vomiting produces some aerosols.

How much that uh contributes to spreading of disease we don't know. And s similarly that there was a case report during SARS of flush toilets and air from the uh the sewer system spreading through an entire uh apartment complex. Similar kind of idea that it instead of being spreaded randomly around the apartments in the entire complex, it went up in the direction that the the hot air plume would have gone, and so you can infer that maybe it did go through aerosols, but it's not the greatest level of evidence by any means. Yeah, that makes a lot of sense. And so I think the next question that tags onto that nicely is what exactly happens to aerosols after they're expelled?

Do they kind of travel f you know wide and far or do they land pretty close to proximity of the original event? Yeah, and this is another one that is almost impossible to to to answer. I I've got stuck in these this series of papers that if anybody's a math geek, you should definitely download some of these, because they have these formulas that will go on for page after page trying to predict where these things go because All sorts of uh things matter here. It matters the local humidity, the temperature, the airflow, the yeah, you know, just the the temperature of your body. There's all sorts of stuff about turbulent flow.

Um I I I think If you really want to oversimplify it, uh, the bigger a particle is, the closer to you it's likely to land, because the idea is it's heavy and gravity is going to pull it down But once you get down below, again, there's no good cutoff, but maybe twenty microns, uh the buoyant force it's likely to stay in the air, and then the spread is very much dependent on the ventilation of the room and it can travel Much, much further. There's some diagrams here that show clearly that it swirls around a five-meter room multiple times. Maybe the most important thing we should talk about is You know, there's no thing that is going to guarantee where a particle is going to end up.

So this is all going to be based on a probabilistic uh kind of pattern. So even if you say most particles end up around one meter You know, there's gonna be some more and some less. So it's not like there's gonna be a magic force field at any number that you hear that prevents all particles from getting to you. Is there any evidence for this two-meter force field-esque rule I've been hearing about? And I probably need to clarify two meters is six feet.

Yeah, it's a difficult one to answer because I what I don't want from any of this, again, everything we're talking about is a lower level of evidence than probably other almost all of your other podcasts. Uh and I don't want people to freak out. I think the easy the str the strict answer is no, absolutely. There's no force field There's all sorts of studies out there that have measured sneezes that have gone as far as seven or eight meters. Clearly, droplets and aerosols can both travel further than that six feet or two meters.

But uh because this moves through 3D space, um it's it follows that inverse squ GWER law that a lot of people uh know about. Basically, the further away uh you go, it is not just going down linearly, it's going down exponentially. So as as you move as you double the distance away from the source, you're actually going down by a factor of four the number of particles you go and as you quadruple it's it you're going down a factor of sixteen. Um the b bottom line is There is no strick cutoff, but for most patients, even if they're coughing and sneezing, by the time you reach two meters away, the chance of you interacting with one of these particles goes down fairly dramatically.

So ideally I would always, you know, take off my p PPE behind a screen or a door. But for the average patient, uh uh in is as an emergency physician right now, we're seeing tons of these patients, and I'm swabbing a lot of these patients and I'm interacting with them, and I'm telling you, I'm pretty comfortable. I do take off my PPE uh unless the patient's critically ill t uh two or three meters away from the patient on a regular basis. And I know that the risk isn't zero because it's it that's impossible, but it's pretty damn low and I'm I'm comfortable with that. And I think one of the keys that you've kind of consistently mentioned is that the distribution is probabilistic and the mathematical models that I'm that are being referenced on your website all relate to the probability that a droplet will fall in a certain area.

And I'm looking at the aerosol trajectories from this paper by Chen in 2010. And again, all of it's related to the location that they model that the vast majority of the droplets will travel. But that doesn't necessarily mean that all droplets are going to fall within that uniform space. Yeah, I I think this tells us to be careful, but also not to overreact. Uh we may come back to COVID specifically, but you know, there's a chance that it's airborne, but we know for sure that these big droplets that spray all over the patient are just much bigger.

They contain much more virus. you're still a lot more likely to catch the uh the virus when you're within one meter of the patient than you are, even if it's technically possible that you could catch it through an o airborne route at three or four meters. Gotcha. And I think that's a neg nice segue into million dollar question, uh does coronavirus travel by droplet or airborne? Because I would say that a lot of the concern over surgical masks, N95 usage, all centers around this conversation. Yeah, and and maybe we should have put a caveat right at the beginning.

Um I love reading evidence and I read a lot of studies, uh, and so I'm comfortable reading these studies. But you gotta be really careful when you're wading into a subject matter that you don't have a lot of expertise in, right? I I I've now spent a, you know, two to three weeks trying to become an expert in this, but You know, the infectious control people have done it for for decades. Um so it is really tricky trying to sort this out and th you you hear strong opinions on both sides. So you you have infectious control people saying one hundred percent This does not spread by airborne and other people who are saying look at the evidence, it does.

Um and so when when that comes up, I I I don't know how to interp interpret the expert opinion. Uh there's pretty scanty evidence here, uh, but I think there's enough evidence uh that I would say there's at least a chance that SARS uh spread through airborne. And I think maybe the strongest evidence out of a lot, there's a lot of these case reports, we mentioned a few of them, where if you look at the pattern of spread, you know, you you look at an entire apartment block, and if it was by touching the elevator buttons or whatever, and then spreading by contact spread, it should be a very random distribution, but when you watch it rise in a specific drive, it follows the pattern of ventilation.

Uh That sounds a lot more like airborne spread. And there was a couple uh case reports in hospitals of of the same kind of thing where it follows the pattern of ventilation. That's still relatively uh weak. To me, the strongest thing was um let me see if I can find the actual s uh study here, but the the what there was one study out of Singapore where the strongest uh predictor of um Whether you caught SARS in the hospital was the ventilation on the wards. Um and ventilation really shouldn't have anything to do with droplet or contact spread.

Uh wards that had poor ventilation had higher rates of transmission and uh wards that had uh good ventilation had much lower rates. So I I think we can say SARS pr probably had it at least possibly spread by an airborne route. That's a lot of a lot of provisos in there. And then if you look at COVID specifically, it's only been around for three months, but they've been doing swabs in places where droplets should not go. So up into the ventilation system, uh outside patients' rooms, and they are finding um the DNA in places that could only be reached by aerosol droplets.

So again, I I think the best answer is There there's definitely a possibility of aerosol. At the same time, I would think for every one airborne spread, there's at least a hundred, uh, if not more, who are spread by droplet or contact. So now that we've covered some of the evidence and clearly a lot of this is very weak circumstantial type evidence. What do we know about comparing N95 and surgical masks? Because I've actually seen a lot of people post that uh one review they did where they looked at influenza and they compared N95 and surgical masks in health workers and they healthcare workers and they use that as um I guess ammunition to support the the hypothesis that surgical masks are just as good.

Can you tell me more about that? Yeah, and this is one of those embarrassing examples where your own biases have a lot to do with how you interpret the research because I have seen these these RCDs have been coming out for years. And N95s are always harder to find and less comfortable. And so every time I see one of these RCTs, they're like, oh, the surgical mash is just as good for my influenza patients. For the last five years, I've sort of been like, yeah, that's great, I'll wear a surgical mask.

Uh but now that it matters maybe a little bit more, uh, I actually took the time to really dive in and read some of these articles. Um and all I would say is it is true that uh thus far the trials have been relatively uh negative, haven't shown a difference. But I think we need to really understand the methodology to be sure that that is a uh a a true finding of no difference. Um and at this point I I think you would you have to say that it it that we can't be certain at all. So the big uh systematic review that everybody's been talking about is long 2020 in the Journal of Evidence-Based Medicine.

That's the one that came out this year. Um And there were there were six uh RCTs included, all in inf influenza, as you say. So step number one, we have to be sure that influenza spreads in the same way as COVID does. Uh and then when you start digging into it, you have you realize that people aren't as um good at wearing their N ninety fives as they're are wearing surgical masks. N ninety fives aren't as comfortable.

And so if you're wearing your N ninety five less often That just biases the study to have a null outcome, uh, which may not be true if you were wearing your N95 all the time, which I'm pretty sure in COVID we're all pretty happy to be doing. Uh the studies also didn't measure other really important confounders like how often did you wash your uh hands, uh how long did you spend in with sick patients? Often these happened in outpatient settings or just on wards, which may not translate very well into the ICU settings or or you know when we're intubating, which is these those are the things that we really, really care about in in COVID.

So there's A bit of a concern about internal validity, uh seeing as um you know if you're not wearing your N95 mask, it doesn't matter how much better it is, and there's definitely some concerns about external uh validity as well. And then if you finally dig down into it Even there's actually I I think more than 9,000 patients in this one a systematic RCT uh systematic uh review, which sounds great, but Sample size is not the total number of patients you have in a in a trial. It's the number of patients with the event that you care about. And overall, not that many healthcare uh workers actually get uh sick, especially if they're wearing masks.

So in the end, the N the the number of people who actually get sick is quite small. And so the error bars are massive. And if you look into it all of the uh point estimates are on the sides of N ninety fives being better. Uh uh you know, I look for for five studies that looked at influenza-like illness, the relative risk is zero point six one. So favors N ninety fives, but the confidence interval goes ever everywhere from And so the only fair conclusion is we just don't know I think you did a really nice job there kind of highlighting some of the tricky aspects of evaluating evidence because I've seen many times people will look at the abstract for a paper take the conclusion at face value and use that to support the argument.

But the reality is there's a lot of problems with, like you were saying, internal and external validity. And in some of the case reports that I was even reading relating to the what you said about N ninety fives. The a a lot of these pati a lot of these um healthcare workers who ended up getting transmission of of these diseases they were studying weren't even fit tested. And so if you'd never bothered to read through the case report to learn about the actual protocol, you would assume that They had just gotten this disease, but in reality they were not fit tested, so they were effectively useless. Although that might fit pretty well with real life for a lot of people.

Yeah, I completely agree. All right, Justin. So um how about a obviously this is going to be very difficult, but how about a quick summary of what we just talked about with respect to aerosols, and then we'll head into the evidence about aerosol generating procedures. Yeah, I think maybe the most important thing I can say is that there is just no really strong science overall. And so we need to be very careful making definitive statements here.

I think the second thing that I think is really important is the more I read this, even though I think that COVID probably can be spread by aerosol or or by airborne route. I actually think I'm pretty reassured by this data because it makes it much less likely. The further away you get from a patient, the more diffuse the air is, and most of our hospitals have really good vent ventilation. Uh and so even though it's technically possible, it's also unlikely, uh which is good. So I think really good hand washing, really good eye coverage, that drop the percussion is going to get us most of the way there.

And then the idea that in things that are really gonna stir up aerosols, lots and lots of extra droplets uh that we'll come back to with these aerosol generating procedures, that's when you really got to care about that N95 and uh keeping yourself really, really protected And then I guess the the final point when you put put all this uh together. Um when you when you try to piece everything together, yes small droplets, these aerosols travel further. Large droplets are going to drop near the patient. The honest truth is we don't know everything. Th there's gonna be a probabilistic spread everywhere.

Probably the most important thing that we're all have all been talking about is that idea of the precautionary principle. So We do not know enough, and the evidence that we should be looking for at this point is not evidence that something is harmful, but evidence that it's safe Uh so before uh getting rid of levels of PPE, we really, really should be looking for strong evidence that it's safe. Now that's hard in a time of pandemic, and it's hard in a time of PPE shortages, uh, but that's what we should be striving for at very least, I think. That's awesome. I think it's a really good summary and you hit on a lot of key points and big takeaways from that first part of the discussion So the second part of the discussion that we'll move to is aerosol generating procedures, and this is off of part two of your your series.

Yeah, and so I I think as I said in part one, I think maybe one of the biggest things is we use aerosol generating procedures to mean two different things uh that we would would really like to try to distinguish between. procedures that actually produce error saws, i. e. they uh convert droplet transmission into airborne transmission, and then just procedures that are really high risk. procedures where there's a high chance that we're gonna get sick. Unfortunately, no studies that I found were able to distinguish those two. So we just have to we have to take them together, even though I do think there are two different and important concepts Also, like we mentioned in part one, the evidence is all incredibly low level.

There's a systematic review that sort of drives all of this, which is uh by Tran. And if you look at their like grade table, every single study listed has a very low level uh on the the grade level of evidence. uh so that's not a great place to to start it with we're mostly looking at retrospective either case control or cohort st uh studies when we're trying to figure out where the risk of transmission is. I think when you're looking at those retrospective studies, you also have to realize that almost all of these procedures are impossible to pick apart from each other, right? So You don't bag valve mask somebody unless you're in the realm where you're gonna do some airway management.

So how can you tell whether the the person caught SARS while they were intubating or while they were uh giving some VVM ventilations? You just don't know And the other confounder that's tied right in with that is we don't do any of these procedures unless the person's really sick, and really sick patients are likely to have much higher viral loads So even if we decided to step back and say, oh, you know what, I'm not going to use non-invasive ventilation, that patient may still have a very high viral load and they still may be at really high risk, uh, because that may be where the transmission is uh coming from. Um there are probably some other things that we can go through as we get into these individual uh datas.

I I think the the biggest and most important thing is uh again, we just don't have great uh data people have not measured the aerosols them themselves. We have a few studies that look at you know how far your breathing goes and then we have a few retrospective studies that look at what your chance of catching and they're basically all in SARS um what your chance of catching SARS is if you did certain uh procedures. So we can go through it. But at the end of the day, we're gonna say I don't know a lot. So I think the f the first place we'll start uh is with intubation.

And um like you said, a lot of these things are impossible to tease out from one another. And the reality is a lot of these procedures that we do are all bundled together. Um and so keeping that in mind, do we have evidence that there's an increased risk of transmission to healthcare workers with intubation? Yeah, intubation must be the easiest one of ox. I don't think anybody's thinking about uh intubating anybody without full airborne precautions So the the systematic review uh again by Tran has highlights these in divides it into two different groups, the the cohort studies and the case control studies, but The results are the same.

We get an odds ratio of of six point six in in both of those. Uh so pretty good evidence that the people who caught SARS were um or or intubating was associated with a higher risk of catching SARS among the healthcare workers uh who were working during that time. Any specific things we need to think about in terms of the PP that these providers or clinicians were wearing? Yeah, not all of them said, but I I highlighted the study by Fowler uh specifically, uh, because they did find so they found a relative risk of uh 13 um so I think 130% uh here. Uh and they specifically said that all of their staff in their study were wearing full PPE, which included N95s. although they note that not everybody was wearing eye protection, which complicates uh a little bit.

Um so it's a little bit concerning because we'd like to think that uh full PPE would really eliminate a lot of this aerosol generating procedure risk. And I'm not sure that we know that it does. On the other hand, uh like you said earlier, uh uh when SARS snuck up on people, a lot of people were not properly fit tested for a ninety-five mask, which they don't talk about here. And if you're not wearing eye protection, uh And you're getting straight up into uh people's mouths and in two thousand three there was not a lot of video laryngoscopes, uh you can imagine a droplet might have been a big uh big issue. Yeah, yeah, that's a really great point because right now we're doing video laryngoscopy glidescopes and it's all indirect, so we're at least we try to be as far away from the airway as possible.

I guess the only other really interesting thing to me about um intubation that's worth noting is going back to that basic science. So aerosols are produced when you have high flows of uh of air over top of some kind of liquid or or mucus That doesn't happen when you're intubating. Uh so I'm not actually sure that it produces aerosols as much as it is that we're right in touch with where the virus is. And so it's a super high risk of transmission. Uh but I'm not sure that The aerosol or airborne spread is a huge part of that, unless I'm overlooking uh some of the mechanics here.

Um, and maybe when you go back to the BVM, but the actual act of the ringoscope itself seems like it would be a completely different concept. Based on this evidence, and like you said, some of it is at fairly high risk of bias, it seems like uh intubation is a very high-risk procedure. Absolutely. The next procedure that I think is thought of as extremely high risk is chest compressions. And I know this is not something that people want to think about with respect to their patients, but Do you remember reading an article about uh a superseding event in which SARS was transmitted to Don't remember the exact number, but maybe around more than three or maybe around five people uh in the course of of CPR on this patient.

So what do we know about chest compressions and and the evidence of transmission? Yeah, this one's an interesting area um because there's actually a fair amount of debate uh coming from the emergency department side of things. Um there were some really intelligent um and well-meeting infectious disease uh people who came together and said specifically that chest compressions are not an aerosol generating procedure, although I couldn't find their science for that. Whereas DHA and IlCOR have all come out and said we think that it is, although they admit very rightly that there is not very good evidence. Um if you go to transmission, um everything's based off this uh trans study um and is all SARS patients, uh, there is one case control study that showed an increased risk, a statistically significant increased risk. with an odds ratio of 4.52.

There were a couple cohort studies where if you combine them together the risk isn't uh increased. Uh I think you gotta be careful there because Picking up the individual uh studies, you know, you find that one of the studies has an odds ratio of 0.4, as if chest compressions were somehow going to be protective. Doing chest compressions means that you're less likely to get SARS. Uh and if you actually just dig into that paper, there was only three Three nurses. Three nurses did chest compressions and didn't get SARS.

Uh and then that gets published with the lower odd odds ratio. It tells you what kind of numbers we're dealing with and what we're trying, we're trying to make conclusions off of uh data that is just not good enough to make conclusions uh often. That's right. I don't remember the last time I saw a study with an N of three. And then of course if you're doing chest compressions, you you're never just doing chest compressions.

So at the very least you're bagging uh maybe you had an LMA. Most likely a lot of these patients got intubated, which we already said is super high risk. Um, so I'm not sure I'm I'm not sure. I personally would treat chest compressions as if they were aerosol generating, as if they're really high high risk. But it it maybe is a bit of a mute point because if they aren't, something else you're going to be doing at the exact same time uh in the airway management realm will be uh and so clearly we we need to be protected during these uh resuscitations.

I think the next procedure which really ties into both CPR and intubation is bag valve mass ventilation. I know Lots of our guidelines nowadays are recommending that you do rapid sequence induction straight to videolaryngoscopy and avoiding bag valve mass ventilation, of course with the claim that bagging a patient will generate aerosols. What's the evidence for that? Yes, this one's uh uh all of these I guess are pretty controversial. So there's only a single study that looked at the risk of transmission uh here.

Uh it was statistically significant. As compared to the odds ratio of 6.6 that we saw with intubation, the odds ratio here is only 2.8, so lower, but statistically significant. Interestingly, in the same study, so those are people who had bag valve mask prior to intubation. If you had bag valve mask ventilation alone and the patient didn't go on to get intubated, Uh they didn't find it as statistically significant increased risk at least. The odds ratio is one point three with a pretty wide confidence interval.

I have been thinking about that for a while. I can't imagine why anybody was bagged during SARS without going on to be intubated unless they had a I guess an end-of-life discussion would be the only thing that I can think of. So it's a bit of a weird clinical scenario and I don't know what led to that distinction, but it is a little bit funny that the risk of BVM is associated only with the people who got intubated and not those who didn't get uh get intubated. Maybe the most important thing to think about here is we we we're talking about the risks uh a lot Uh but I think you need to consider the flip side of the coin, right?

So you could definitely avoid BVM ventilations because there's an increased risk here. And on one side of it, that sounds like it makes a great deal of sense On the other hand, you have to ask yourself what you're going to do instead. And really the only alternative here is what you're saying is we're going to rush straight into intubation. But if we know intubation is super high risk. And yeah, we're gonna do it anyway, but which is gonna be higher risk intubation?

The one that's controlled with the patient who has SATS higher than 96% and you have all the time in the world, or the one that you're rushing because the sats are all already falling and you haven't taken the time to just vent the patient up I think there probably is a role for gentle ventilations with a good two-hand seal, hopefully with waveform capnography, so you can see that you don't have a leak I I think there's probably a role for it, but I think the evidence shows that there is going to be an increased um uh transmission. And if you go to the basic science side of things, I guess we should mention as well It be BVM is so hard to assess because people do it.

There are some people who are really, really good at this, and there are people who are not so good. Oh, absolutely. And there's a wide range of spray around that BVM mask depending on who is doing the bagging. And I think this relates to a topic we had covered in the past, which is the topic of mask ventilation and when is it actually difficult? And What we had concluded was that even though the ASA has a definition of what difficult mass ventilation is, when you actually operationalize it, it's very tough to say what exactly difficult is.

And so I think one of the studies you might be talking about is the one where they compare the adequacy of the seal with bagging between different types of clinicians. So nurses Anesthesiologists and respiratory techs. Yeah, there's a couple different studies by Chan. There's a 2013 and a 2018. I I don't remember the one that included nurses.

It's all a bit of a m sketchy surrogate because all they put smoke in the uh this a simulated patient's lungs and then uh use some lasers to see exactly how far the smoke goes. Um, you might be reassured because the smoke only goes about 30 centimeters away from the patient. So that doesn't actually sound like significant spread. It's not like you're gonna be holding your face 30 centimeters uh fr from the BVM, you should be further away from that uh most of the time. But I think if there's a leak that we're seeing Yeah, they worry would be is that extra pressure turning droplets into aerosols, in which case they're gonna float a lot further than that that 30 centimeters.

Again, this is a level of data that we're we're dealing with, and I th I think the answer is I I just don't know. I guess you could put a flip fly to what I said originally. I said, you know, the only other option would be go to intubation, but actually maybe a much better option is just to go to an LMA. Uh and then that eliminates sort of both of of your options um even though that may not be perfect either, but I don't know So the next thing that I think we is worth talking about is if we don't go to intubation, what are our options? And I think there's been a lot of discussion about non-invasive ventilation and All the recommendations I'm currently seeing are to avoid non-invasive ventilation, again, because of the claim that these methods of oxygen delivery are gonna either generate aerosols or increase risk of transmission or both.

Yeah, and there's some really fascinating discussions going on right now because that's definitely the guidance we have as well. But if you listen to the people who have gone through this in Italy and now New York City, a lot of them are saying that's probably a mistake because by avoiding this really important tool you end up putting everybody on ventilators. And we know that patients' lungs don't enjoy being on ventilators, but also you use up your most precious resource. And so I think a lot of people who have been hit the hardest are actually making the argument that Even though there's maybe an increased risk to healthcare workers, uh, this is something that we need to consider.

Uh and so trying to grapple with the data is really, really important. Like everything else, it is incredibly weak. There were two studies uh that When combined, show an odds ratio of 3.1, i. e. there's an association between patients being on non-invasive and healthcare workers catching SARS. Um the basic science here might actually be a lot more fascinating and honestly I haven't been able to get through all of it yet because the non-invasive is not all made the same. There are many, many different kinds of masks.

Uh I don't think anybody's measured those uh They look like the old school uh underwater scuba things where it's the whole helmet and goes over your entire head. Uh you're using a lot of those in Italy, but I don't think they've been studied at all for this. But there's the full face masks, there there's different exhalation ports, there's the nasal CPAPs. Um and so I found one study by that about a specific face mask, the Quattro Air face mask, that basically didn't show any significant air dispersion using the same technique we saw be uh before. Basic basically they couldn't measure any of the smoke spreading away from the patient.

Whereas if you use the normal masks or nasal masks, you can easily see smoke spreading 20, 30, up to a meter away from the patient. So it probably is much more nuanced than just non-invasive, not non-invasive, but the different techniques and the amount of pressure and the flows that you're using are probably important as well. You mentioned that the quattro air is a very unique design that continually vents air, and I think that just speaks to how subtle that is because all these models are designed differently, they fit differently on different patients' faces. There's a lot of inherent risk with interpreting data from that kind of uh information. Yeah, and none of them have actually been studied.

I mean it's not hard to search uh the internet right now and see people jury rigging designs where you have a viral filter the first thing after the CPAP mask. Uh which makes a whole heck of a lot of sense, uh which would essentially eliminate the risk from normal use, uh and logically it makes sense, but if you get any leak around the mask uh then you're in trouble and unfortunately there's some great studies showing very, very little leak. Again that quattro error uh one that I talked about uh showed no leak around the mask in their study, but it was on a mannequin. So the mannequin's not moving and flipping over in their ICU bed and they're trying not trying to prone it and things like that.

So it it's really hard to extrapolate. I guess the other thing that's really fascinating to me here. So Uh the odds ratio here is three. I told you the odds ratio for intubation was six. So you might look at the face of it and say, why the heck would you go to intubation, which is a higher risk? for something that is uh to to avoid something that is lower risk.

But then you gotta factor in the time factor. And I was saying telling you before my wife is an ICU nurse. So we can Do a high-risk procedure that takes ten minutes right at the beginning of the person's stay, or we can spend the next week uh subjecting our ICU nurses to twelve hours of shifts a day of somebody on non-invasive. Uh I don't know how that math plays out. It's it's pretty complicated.

Yeah, I think that introduces a whole level of complexity that like we were like we've been talking about, there's just nothing to to be able to allow us to grapple with that question. So I think one of the natural corollaries to non-invasive would be then high flow. What do we have about evidence or basic signs with respect to high flow nasal cannulas? Yeah, because all of our transmission studies happened in the era of SARS, high flow just didn't exist then, so we have none of these same kind of uh transmission studies. There was a really fascinating study uh by Liang in uh twenty nineteen that looked at patients with pr proven gram negative bacteria and they put them on high flow nasal cannula.

Uh it's only twenty patients, but they put basically petri dishes around them and they did some air samples and They couldn't uh culture on any of the petri dishes the actual gram negative bacteria. There was a little bit more bac uh bacteria in general, probably from the nasal uh pharynx. So That's a really interesting study uh saying that the high flow is probably not taking things, uh bacteria at least, out of your lungs and spreading it around their room. It's hard to v uh to put use that data in COVID though because the high flow clearly puts really high flows through your knee your your nails. Uh and Viral pneumonias tend to have much more upper respiratory tract in involvement.

So I I don't know that that bacterial pneumonia study really applies very well in our current uh setting. What we do know is if you put a lot of flow through people's uh nose and then they cough or breathe, those smoke studies, you can see the spread going much, much further than any of the other uh modalities that we've talked about uh so far. And I saw this reference on your website and something that I had also seen, which is the discussion by uh one of the these manufacturers who Are showing that even though high flow might increase the dispersion of particles, if you put a surgical mask over the high flow nasal cannula, then that risk is mitigated.

What do you think about that? Yeah, that 's a tricky study because it's only, as far as I can tell, published on the manufacturer's website, which is always a little bit of a red flag. Uh I think it goes back to the basic science that we talked about before. The difference between droplets and aerosols. So what they're measuring and what we keep talking about is how far this smoke gets away from the patient or how far people breathe.

That really matters for spreading of droplets. Like, you know, you have that powerful sneeze and you can see snot shoot like six meters away from you. Uh if you put a mask on, that snot won't get six meters away from you However, air still goes out around both sides of the mask. And the really, really small aerosols don't hit the mask because they're so small, they just go wherever the air goes. They follow the air uh channels.

So if this is creating aerosols, uh I don't know that that data would will tell us anything. It only tell us tells us that the droplets won't spread really, really far, which is good. Uh but what we really want to know is is it converting something that should be droplet into aerosol transmission? Uh and I don't think that we we know. Um there is an interesting argument also made uh I believe by manufacturers that uh aerosols are more likely to form or really small droplets are more likely to form w when it's really dry.

And so if you humidify the air You're just gonna get lots more big uh big droplets, uh, which might help. But again, that 's getting pretty theoretical. So it sounds like it's pretty difficult to know with high flow nasal cannula right now with the evidence that we have and a lot of it's gonna be up to clinician judgment Yeah, I I think we're hearing more and more people are using these. I I would personally be comfortable looking after a a patient if I had full PPE and a negative pressure room. Uh the question is what we do when we start running out of of that stuff.

Um and I still don't think we know the risk-benefit of that versus just going straight to uh to an intubation. It's impossible to know from the available available data. So I know in the operating rooms we're often using things like face masks and nasal cannulas, and I would say that the most pervasive claim around those is the need for low flows. What do we have in terms of evidence about transmission related to these oxygen delivery devices? Yeah, I the I have a hard time in interpreting these numbers.

Uh one of the biggest claims that I saw circulated was you know basically if you were below six liters you were fine and above that that's when you started to get aerosol generating uh type stuff. And I I couldn't find a study with uh with that supported that at all. I've heard that exact claim, and it's amazing that that just spreads from country to country apparently. Yeah. I I I have searched so long trying to find anything to either refute or support that and I've got nothing.

During SARS, the use of an oxygen mask was associated with transmission of SARS. Um there was There was one study where it was the highest risk factor overall. So there was an Oz ratio of seventeen, although the con the 95% confidence interval goes everywhere from 1.8 up to 165. So IUB we have nobody clue. Uh and then there was two other studies um that were uh cohort studies that again the odds ratio is four point six.

Uh people have pointed uh at this and actually said Uh there's no statistical difference, therefore there's no risk, uh, which I think is a little bit silly because the point estimate of 4.6 seems pretty concerning, but again, massive confidence intervals, and so I think we just don't know. So uh this is one of those ones where I I think those confounders really, really matter. We don't put people on oxygen who aren't really sick. So to me, this might just be a marker of really sick patients. This might be a a marker of higher viral loads.

Maybe I'll mention two other things that were highly related to transmission of disease, but we don't think of as aerosol generating procedure. There was one study that showed that um putting in a fully catheter had a very high risk of contracting SARS, um, so the odds ratio of over five. And another study that just being in the room while somebody had an ECG done. uh or putting a peripheral IV in had a high risk of contracting SARS. I think what those tell you is that spending lots of time in a room with a really sick patient uh puts you at risk one way or or another. Um I'll throw out one more number, even way too many numbers here, but we've talked a number of times about inbate uh intubation has an odds ratio of 6.6.

Right. And that sounds pretty high risk. But from that same uh but there's one study that looked at a number of different factors uh found that intubation in that study actually only had an Oz ratio of 2.79. But an Apache score greater than 19 had an odds ratio of 17. And having that classic PAO2 to FIO2 ratio of less than 60 had an odds ratio of uh almost nine, so way more than intubation.

So again, I I think Being really, really sick and spending lots of time with sick patients is the thing that's probably high risk. I'm not sure what to make of the the oxygen masks in and of themselves. Uh the more I look, the more I do find that you know, again, these smoke studies, if you the higher you turn up the uh oxygen, it just makes sense the further things spread around the room. Uh so I think the lowest flow you can get away with makes sense. We we honestly don't have all that much use.

Like you should never be giving patients oxygen who don't need oxygen. Right. Uh that's a pretty easy rule rule to go by. And if they need oxygen, I guess the only dis uh the only decision we have to make is how quickly you intubate them versus leaving them on that uh filtered non-rebreather for 12 hours and and watching them and uh ever everybody hospital seems to be coming up with slightly different uh approaches to that at this point. I think one of the big takeaways as you've mentioned is that patients requiring oxygen are probably sicker in general, and that means they Again, this is a claim that might need to be investigated, but they might have higher viral loads, so these devices are creating conditions for aerosol generation.

And ultimately a careful risk-benefit analysis needs to be conducted by the clinician. Yeah, I I I think that's really our only approach. I wish I could have joined you. I I've listened to some of your other podcasts and I I generally like to be able to dive into some really good data and come And a as far as I can tell that just doesn't exist uh right now, which is unfortunate because this is a uh higher stakes uh decision than a lot of the things that we talk about like you know running uh pressors through a peripheral IV or whether somebody should be NPO or not, right? Like it is the thing that we have the weakest data on.

It's it's a little bit frustrating I completely agree. It's amazing how sometimes with the things that matter the most we have the least to to support our claims about. We have two more to go through. The first is suctioning, and then the second is bronchoscopy. I think broncoscopy we're trying to do as little as possible, but suctioning It's hard to get away with, especially when we're excavating patients.

They have to be suctioned. I imagine up in the unit when they're being extivated, they use the inline suction, but we just don't have that in the operating rooms right now. So Um what do we know about the evidence of transmission with suctioning? Yeah, so suctioning is fascinating uh because if you just go to common sense and uh a little bit of basic science, uh it actually realistically could be argued to decrease your risk by a fair amount because you're getting rid of all of the mucus and droplets and everything that could spread it. And even if you had good enough suctioning, you could actually Suction the air.

That's what a negative pressure room is, right? So if you use one of those silly intubation boxes that you see circulating the internet and put a bunch of suction in there, uh you could presumably decrease airborne trans transmission. On the other hand What causes aerosols is air flow. So if you have the suction in the patient's mouth and the that's causing air to flow out of the lungs, you could be pre creating a bunch of little aerosols uh coming out of the lungs. And so it's really, really hard to to tell.

Uh I wish I could give you a bit a stronger answer here. Again, there is uh evidence of transmission in SARS. So there was an odds ratio of thirteen point eight in the uh in the one study that uh Sorry, uh I'm reading one study looked at it, that's Loeb 2004. When you combine things together in the uh meta-analysis, uh the odds ratio was only 3.5 and it was not statistically uh significant. for suctioning before intubation. And actually the suctioning after intubation doesn't look too bad.

Odds ratio 1.3, again, pretty big confidence intervals, although not as big as we have seen in other places, 0.5 to 3.4. The the honest truth is so suctioning at the time of intubation is you have no choice. Uh you you need to be able to see where you're putting putting the tube. Uh You're just gonna suction, there's no choice. Uh if you have the option, clearly suctioning in a closed system makes more sense than not Uh I think suctioning for in your uh situation, and actually one procedure that is isn't on my list because we don't do it in the emergency department, but I also didn't come across anywhere is just the idea of extubation, pulling a mucusy tube out of somebody's trachea uh seems incredibly high risk, but I didn't find any data on it at all.

Uh I I think you just gotta go back to some common sense. I I would treat it as aerosol generating. I I to me that Extubation is probably just as high risk as intubation. Uh however, um the suctioning component just makes sense. If I if I can limit the amount of mucus Flying out when I pull a pullback tube, that that makes some sense to uh sense to me, but uh science is very limited Speaking of limited science, how about bronchoscopy?

It looks like this is even worse in terms of evidence. Yeah, there was uh I I think two studies that mentioned it, but uh slightly high increased odds ratio at 1.9, but the odds ratio goes all the way from 0.2 to 14.2, i i. e. just massive because there's just not enough patience here. Um I don't ever do bronchoscopy so I I can't tell you what the benefit is. Everything here is a risk benefit. So I th nobody should be saying anything definitively like you cannot do this procedure, but at this time during COVID, we should not be doing procedures that aren't necessary.

And I guess that's the decision that our intensivists have to be making. Awesome. I think that was a really solid review. In spite of the incredibly limited evidence, I think you did an absolutely amazing job, at least looking for what is currently out there and I think one of the key things is just thinking about the data and evidence in a way that you can frame it appropriately. And I think You know, you included this in your discussion section, and I'd love for you to just go over some of the insights.

And I think that a lot of this is just going to be clinician judgment, but I think the way you have Talked about applying the evidence makes a lot of sense to me. So I think it'd be awesome to hear that. Yeah, again, I I think the most important thing that I've been preaching over and over again is simply don't tr go out there making definitive statements Because we just don't know. And then we need to fall back on that idea of the precautionary principle. If we don't know, we need to make sure that we keep ourselves uh safe.

Uh I think when you look at this data, uh maybe the most important lesson over and all over is Uh maybe less that it than it's the procedures, um and more that it's really sick patients that are the the high risk to us. Um I think out of all the quote unquote aerosol generating procedures or th or things in these studies, the fact that the transmission rate is the highest just for uh the people with the high Apache scores or while you're putting in a urinary catheter, I think is really telling. And so for me uh it means that obviously well intubating uh I wear full airborne precautions. But if I have a patient that's still on the edge and I'm trying to decide whether they need intubation now or whether I can get away with some something else, but they're coughing up a storm and their sats are you know, eighty-eight percent and they're starting to require more oxygen, uh or maybe I have to spend some time in the room and put it in a central line, what whatever it is.

I I think you can most of us can use some clinical judgment and decide that the the really sick patients are probably the ones that we need to use the highest level of precaution on. Um and I guess the other major point that I said here is uh way too many of us are treating these as uh I guess I call them one-sided decisions. Uh we are absolutely going to avoid uh BVM ventilations because there's a risk associated to it. But we also have to consider the other side of that coin, that the risk associated with not doing things. I because I I I don't know.

I don't know what the which is higher risk. Uh BVM ventilations or that rushed intubation uh that is always uh just way more chaotic and and much more messy. Uh is it higher risk to keep somebody on CPAP for three days or to do that again that that highest risk procedure which is um intubation. You know, do we do I just bag somebody for, you know, I'm Most of these arrests are pretty short uh running. So if I have a good seal, I could bag them or I could put it in LMA or I could intubate them.

I these are tough, tough decisions. I I think reading through some of these uh tells you that there's a lot of gray area uh and that you need to have good discussions with your department uh so a you don't have to make the decision in the moment when you're already stressed out. And B to realize that we should be watching for new data and be willing to change our practices as we learn new things from our colleagues who are all going through this right now. I think you made a ton of incredibly, incredibly insightful points there. And I just want to emphasize, I think that one of the difficulties with a lot of this data is that it's just simply not good enough for us to make definitive dogmatic statements about how we should approach things.

Everything really exists along a continuum and even the strongest evidence sometimes has limitations in terms of not only its internal but external validity, which patient populations to apply it to. And I think when we're in a situation where the evidence is so scattered and so weak, it's even more important to be cognizant of that. Yeah, I I think again, one of the worst parts about working right now is the really high levels of anxiety that everybody has. Uh and In my experience at least, the wor the worst part of that anxiety is hearing five different people are getting five different emails on back-to-back days that all say something completely different. but none of them reference science or their reasoning.

And so I think when we're having these discussions, we need to to discuss the major limitations. We need to uh take a time to to explain why. Uh if you're suggesting LMA in a cardiac arrest, that's great. Discuss why you think that and so that as new information comes up Justin, I just want to thank you again for taking the time to to come on the show. I read through your entire two parts on on this subject and I just thought it was so well done that this is content that I think is incredibly important for anybody. in healthcare to be listening right now, not only because of the evidence portion, but really thinking about how to apply evidence and and approach it as a framework.

Thanks for having me, man. It's always a pleasure to try to dispel some dogma, although unfortunately I I don't know that we did a very good job this time around. I want to wish you the best. I think we'll definitely stay in touch and be safe out there. I think times are absolutely Crazy right now, but um I hope you can take care of yourself and your family.

Yeah, thanks. You too, mate. I I know everybody's gonna get through this in in the end and we're gonna come out stronger on the on the far side of this, but uh just keep your head down and look after Look out for yourself, your families, your friends, uh colleagues for the time being. Cheers, mate. To all our listeners, thanks for joining us on the show today.

I hope this was helpful, especially in these stressful times. I do want to finish off the show by sharing some phrases from an email sent to us by one of my co-residents He says, We took oaths of a higher calling, and only a small group of people on the planet is as qualified as we are to do good right now. It won't be long that we're looking back on COVID as a brief brief chapter of our lives, so we have a unique opportunity now to write this piece of history the way it should be written. I share with you lessons learned mostly from errors of my past. Don't compare yourself to others.

It doesn't matter who appears to be doing more or less. Just give your best and be proud of your own role. Don't worry about fairness, but rather be proud in your fatigue of the important work you're doing, and see the importance of being a follower and team player. Find small victories and notice the people who contributed to them. Forgive yourselves and those above below.

And around us, who will make mistakes as we all learn how to make the best of an impossible scenario. Lastly, please hold me accountable to the same. Before we know it, this will pass. Thanks again. and I'll see you guys later. Stay hungry and keep asking questions