Dr. Bonnie Snyder: Small Adult BVM in OHCA (9/27/24)

And I am going to get started. So here we are guys. It is September 27th. It's another Friday. This is the Florida EMS webinar where myself and Tanner Smita we bring on amazing guests who write essentially amazing papers. So what we do is we crawl the literature and then we randomly email people who we've never met before and we say pretty please, will you come and talk about your paper? And today is going to be a great one. We've gotten so much feedback already and she hasn't even started speaking yet. Many people are on Twitter, commenting and LinkedIn. This is a topic that's of interest to many of us and we always look towards the folks up in Seattle to do really impressive research to help guide our profession forward. And Dr. Bonnie Snyder used to be up in up in Seattle and many of her colleagues are on here today. And now she's a second year resident in Emergency Medicine at Cincinnati. And so without any further ado, Bonnie, thank you for joining us. Please make any further introductions and take it away.

Absolutely. Thank you so much for having me. So yeah, I am Bonnie Snyder. I'm here to talk about my paper which we titled the association of Small Adult Ventilation Bags with Return of Spontaneous Circulation in Adaptation Hospital Cardiac Arrest. So a little bit more about me. I was actually born and raised in a small town called Wasilla, Alaska where I was briefly a ground paramedic in a 911 system prior to completing my undergrad there in the University of Alaska Anchorage and joining the University of Washington School of Medicine for medical school. Between my first and second year, I had the amazing privilege of joining the Division of EMS team to start working on this research project which continued throughout my medical school. I graduated in June of 2023 and then joined the University of Washington Emergency Med or sorry, University of Cincinnati Emergency Medicine Residency where I get to pretend that I'm still a paramedic on our helicopter. So we will get into the paper. This was not a funded project, but we do disclose that one of our collaborators, Rob Walker, is employ as a biomedical engineer at Stryker Physio Control. This QR code is directly to our published paper, so in case you want to follow along as we walk through it. So to get started, we know that excess minute ventilation can adversely affect hemodynamics in cardiac arrest by increasing intrathoracic pressure, decreasing venous return and negatively impacting cardiac output. We also know that carbon dioxide is a major regulator of cerebral blood flow and that hyperventilation can cause vasoconstruction which can worsen secondary brain injury. Additionally, nearly half of all out of hospital cardiac arrest survivors who survive to 48 hours develop ARDs which we also know is worsened by large tidal volumes.

With all this in mind, the team at Seattle Fire Department changed from a large adult ventilation bag which has a total volume of 1685 milliliters and an expected one handed delivery of around 700 to the small adult bag which has a total volume of 1000ml and is expected to deliver around 450mls per one handed squeeze. This change was made in July of 2017 and is used for all patients in but we were specifically looking at adults with cardiac arrest.

So we set out to assess the relationship between the small bag use and the incidence of ROSC in ohca and we hypothesized that the rates of ROSC would increase after we made this change.

So some of you might be familiar with the very impressive Creative Arrest Registry that the Medic one and division of EMS team manages. All cardiac arrests treated by Seattle Fire Department are prospectively recorded using Stryker LifePak 15 cardiac monitors as well as a black box audio recording device. These data are then input into a post review software which is demonstrated here. We were specifically looking at our in tidal waveforms and the associated ventilation. Extrapolated through this software but also incorporated into this registry are Data from the 911 call, the pre hospital care record and hospital records. And these are all kind of input and reviewed by the SFD Quality Improvement team.

So we defined our study period as 2015 to 2021. During this time 3,252 patients were treated for cardiac arrest. We figured that there would be some time of overlap as we transition to the small bag from the large bag. We excluded about three months when we weren't sure exactly which bag was in service. After this we made a number of exclusions. We excluded children.

We were specifically looking at patients that had an advanced airway, meaning an eye gel or intubation and had care provided by paramedics. So if you only had BLS care, this is sort of a population of patients that may have just gotten like a single AED shock and had rosc.

If you had a dnr, if your care was provided by a different agency and not Seattle fire, you were all excluded. If you got a crike, if your airway was placed during spontaneous circulation while you had a pulse, we were not as interested in that specific population of patients.

And then just sort of for our capnography data, if you did not have any capnography data, you were, you were kind of fell out of this group. And then also if you.

Were had a pulse throughout the time a paramedic was caring for you, you didn't have kind of any of our time periods of interest, so you fell out as well. This gave us 1,994 patients with 48,008 600 minutes of data. There was some data that was not analyzable from an in tidal perspective, and then There was about 15,000 minutes where these patients had a pulse. And we were not specifically looking at those time periods. So that left us with 26,921 minutes of cardiac arrest data to look at. This was split into our cohorts based off of timeframe like we talked about earlier. So we had about two and a half years in the large bag cohort, 663 patients, and then just over four years in the small bag cohort, and understandably about double the amount of patients in minutes.

So, demographics of our study population, on average, everyone was 62 years old. About a third of them were female, almost two thirds arrested at home. Of the patients that arrested prior to EMS arrival, 63% of them received bystander CPR. Of all patients, 21% had a shockable rhythm initially. And on average, our case length for how long these patients were resuscitated was 34 minutes. These characteristics were statistically similar between cohorts, although within the small bag cohort, more people died at home and also more of them received bystander cpr.

So what did we find? Very unexpectedly, we found that the rates of ROSC at the end of EMS care were less in our small bed cohort as compared with our large bed cohort. So this is sort of a dense table, but you can see we had 40% of all patients had ROSC at the end of EMS care, which was defined as either they were declared dead in the field or they crossed the threshold into an emergency department.

And only 33% of the small bag cohort had ROSC at the end of EMS care. That was a statistically significant difference. This just shows that our rhythms were a similar distribution between patient groups. And while survival was not a primary focus of this study, they did show similar trends for survival to admission, discharge and neurologic outcome.

Excuse me. So because this is a cohort study, we calculated an odds ratio. This is basically the 0.74 number, is comparing the two groups and is basically saying that.

Those patients in the small bag cohort as compared to the large bag cohort, have on average a 26% reduction in the odds that they received ROSC at the end of EMS care. We calculated a 95% confidence interval, which is kind of how likely that number is to be accurate and found it to be 0.6 to 0.9. Because this number does not cross one, it is felt to be statistically significant.

This is a forest plot which just demonstrates this. Sorry, I didn't mention was this is our adjusted odds ratio. Our unadjusted odds ratio is very, very similar and had a similar confidence interval. But we wanted to make sure we were accounting for any potential differences between these two groups. So particularly things about the group's characteristics about the groups that would affect their odds of achieving Roscoe. So those specifically were age, their sex, was their arrest witnessed, did they receive bystander cpr and what was their initial rhythm? So all things that are known to affect their ability to get rosc. And after adjusting for potential differences between the two groups, for all of those things, the difference stood. This just shows kind of by those differences, the relative weight of change for favoring the large bag versus favoring the small bag.

So when we got this answer, we were perplexed. We ran the data multiple times. We wanted to make sure we hadn't first made a mistake, that our data was accurate. And then we set out to think about pretty hard as to why this might be despite our hypothesis. So we thought, could this just be because we are including Covid?

We know that our healthcare system, both pre hospitally and in the hospital, as well as just our day to day life changed incredibly. So the first thing we did was just first exclude all of the patients after Covid was identified in King County. So that was defined as February 27, 2020. We took all the patients out after that and the difference was almost exactly the same.

Our odds ratio was 0.75. With a similar confidence interval that did not cross one and was still statistically significant. We then thought, is there something different about this cohort that we haven't accounted or adjusted for. So we added a number of variables that we felt would be meaningful to patient outcome to our adjustment analysis. So in addition to the ones that I mentioned before, we added the etiology of their arrest by upsene characteristics, whether they arrested in public, how long it took from their 911 call to the initiation of CPR by EMS, as well as the placement of advanced airway by EMS and then also their total epinephrine dose. When we added all of these variables to the adjustment, the association remained the same or similar. It moved slightly, but our odds ratio was still 0.79. Although our confidence interval did widen to 0.99, but was still statistically significant.

We then thought, is this just something to do with change over time? We reviewed whether or not there had been any significant cardiac arrest protocol changes, which there hadn't within our study period. But is there just something temporal about this data because it covers many, many years and this is sort of a more granular demonstration of that. Each of these dots represent the proportion of ROSC for all our patients by month. And we compared kind of the slopes of these two lines and this was not statistically different.

So, moving on to our secondary outcomes, we wanted to look at both ventilation rate as well as end tidal values during cardiac arrest. This is a density kind of heat map which consumed many, probably hundreds of hours of my time and something I'm still quite proud of. So, on the left we are showing each 30 second interval of ventilation that we assessed and we're given a ventilation rate. This is moving kind of forward in time, with time zero being the placement of an advanced airway. Again, Igel or ET tube, although it was a very vast majority.