Thank you, Tia, for that great introduction, and thank you to the SCCM conference organizers for giving me the opportunity to talk about this subject which is near and dear to my heart, and thank you to everybody for packing into this small conference room at 4 o'clock on an evening to listen to it. Here are my disclosures. I do have some research funding in the area of physiologic-directed CPR, and as Tia mentioned, I do volunteer for the American Heart Association. So upon completion of this session, I'm hoping that everyone here understands some advantages and disadvantages of potential CPR quality monitors. I'm going to go over, and for those of you who have seen this talk, I'm going to kind of revisit the idea of rescuer-centric versus patient-centric CPR, and then really, I know the talk was for me to pick one monitor, but the idea here is I'm going to totally cheat. There's not really one monitor. We need to be using multimodal approaches, and I'll kind of allude to that. So I want to start this talk with a quote as the story goes from Max Harrywell in 1996 at the Wolf Creek Resuscitation Conference, where he said that performing resuscitation without measuring the effects is like flying a plane without an altimeter. And so if you take nothing from me from the next 20 minutes, just take this. Monitor something. The idea in 2023 that if you're doing CPR and not monitoring some way the quality that you're doing, that's a problem, and that's really the take-home point of this entire talk, that we just have to be monitoring something. So I do have one additional disclosure while this session is entitled Cutting-Edge Research. What I'm going to talk about here is really practical monitors that really should be available for everybody here in the audience, and hopefully there will be something for everyone. So the idea of rescuer-centric versus patient-centric resuscitation, I think, is a good way to think about the monitors, because that's kind of how they break out. So we have a CPR event here, and for a lot of the CPR quality literature that really started in 2005, we talked about the mechanics of CPR, and was the rescuer meeting the targets that were associated with better outcomes? So rate, depth, release velocity, were you coming up full chest recoil to get blood back to the heart, and were you minimizing your interruptions and chest compressions? And through large-scale study, we were able to develop these evidence-based targets, and this is a model where you can do the most good for the most people. The problem is we have outliers who don't necessarily kind of fit the mold, and so what I have been doing and others in this area have really been trying to shift the focus over to actually seeing how the patient is responding to our resuscitation. And some of those targets that we can look at are blood pressure during CPR, end-tidal CO2, NEARs, and I'm actually going to talk a little bit about pulse oximetry, which I think has some usefulness in the future. So I'm going to totally cheat, and I'm not going to pick just one, but in the monitor itself is going to really be based on the location of where you are, and so I'm going to kind of walk us through a case and how the optimal monitor may actually change over time. So let's just pretend you're walking from the convention center up to your hotel, and you actually come across somebody laying on the ground. EMS has arrived. This is a non-intubated patient at this point, and they're starting CPR, and you want to help. So I think in this situation, probably the best monitor out there is actually the CPR quality monitoring defibrillators. They really enabled more universal measurement of CPR quality. They really put this whole science on the map in the early 2000s, and they moved the needle by allowing us to have evidence to support and change our guidelines. There's different underlying technology across different manufacturers. Some use accelerometers, some use force transducers, but they're very applicable to both in and out of hospital cardiac arrest to make sure that we're monitoring our quality. And there's a ton of literature out there now over the last 15 years associating chest compression rates, depths, release velocity. Pediatrics even got into it in 2010, looking at depths associated with survival. And newer, where we're actually looking at combination of metrics and their association with outcomes. So there's a ton of science here. But what's the problem? Well, they're not cheap. Some of these devices can cost upwards of $30,000 if they're actually transport certified. They take some time to get on the chest, and arguably, those first couple minutes of resuscitation may actually be the most important when you're looking for reversible causes. And there's a learning curve to respond to the feedback. Sometimes it doesn't just come natural. Some of the things yelling at you for actually you to change up your technique. And you have to listen to it. So this was work from Donald Edelson in 2008. This is actually a strip from a CPR quality monitoring feedback program. And what I'm circling here for you is actually prompts from the defibrillator to say compress deeper. And what you can see, we do for a little bit on that bottom curve, which is actually chest compression depth. But pretty quickly, we kind of can squirrel and actually not listen to the feedback, which is why it has to tell us multiple times. I call this kind of the intensivist attention span, where we kind of get distracted and move on to something else. So one of the ways to maybe actually keep the attention there is actually assign somebody for that role. And I like this particular study. I know there's some more work on coaches out there. This is one of the first of the use of a CPR coach. It was an EMS quality improvement initiative. And what they actually did was multiple PDSA cycles. They had a dedicated role. And they actually pulled the captain out to actually be a CPR quality monitor. So pay attention to the defibrillator and coach the team. And this is looking, this is a control chart of depth. They actually saw improvements in other metrics as well. The purple line there right in the middle is when they actually put the coach in place. And over time, what they actually were able to show is they had a 10% increase, 70 to 80% in the compliance for adequate depth. So I thought this was a great study that kind of showed how a coach can actually improve quality. The other real problem with these devices is they don't necessarily, like I said before, apply to all patients the feedback. So this is a study that kind of was the first to really show the heterogeneity across individual patients and how they respond to the individual metrics. So this was an out-of-hospital cardiac, sorry, an out-and-in hospital cardiac arrest study. Very small, 39 patients, but they looked at 42,000 compressions. And what I'm showing you there on the left-hand side is depth of compression versus the actual diastolic blood pressure attained. And kind of in that guideline approach, if you look at the overall average, as you push deeper, the blood pressures got better. And if you're this particular patient, that's fantastic. But if you happen to be this patient, you're probably going to say, stop helping me. And so if we're not actually paying attention to the response, we can actually miss the forest for the trees here. So I'm going to now move on, now that we've kind of got EMS there and they decide to put a breathing tube in. I know that's a controversial subject, so I'm not going to get into that. But let's decide they decide to intubate. And so now I'm really going to move on to some of the quality monitors that deal with more physiology. In particular, this has been something that the American Heart has talked about since 2013, the consensus statement authored by Pete Meany, where they talk about monitoring the patient's response to the resuscitation using end-title or coronary or diastolic pressure. So I'm going to start with the end-title, maybe. So I'd like to go back to one of the classics. There's been a lot of work on end-title and its association with outcomes in cardiac arrest. But I'd like to show this one from Sanders and Chama in 1989, where they looked at 34 patients, both adult in and out of hospital cardiac arrest, and looked at average end-titles between resuscitated and non-resuscitated patients. And what they actually found here is that if all of the survivors had end-titles greater than 10, and kind of as you move with higher end-titles, you get more likely to have survival. This was kind of where that idea that if your end-title never gets above 10, it's probably a futile resuscitation effort. In pediatrics, we've had a little harder time actually getting data to support end-title. This was just published last year out of the Hopkins Group, where they looked at in and out of hospital cardiac arrest using Zoll end-title devices to collect their end-title. And they used median end-title as their predictor. Here is their kind of Utstein diagram. They started with 460-ish total events, and they got down to 143 in almost 2,800 minutes of CPR, where they had both end-title and CPR mechanics data. And while they weren't able to find an association with the current AHA target of 20, they were able to show for the first time in pediatrics that there was a difference between patients who did and did not achieve ROS, about 6 millimeters of mercury difference. This is actually unpublished data. So just to say that again, this is unpublished data that was coming out of the Capcorn Network, funded by the NIH and NICHD, where we actually were able to get a little bit more data on 234 intubated children. And we were able to actually take a look at the AHA target of 20 millimeters of mercury. And encouragingly, for the first time, this data showed that if you actually are able to get your average end-title during the first 10 minutes of CPR at that 20-millimeter target, or even the percentage that you are at that target, they are both highly associated with both survival hospital discharge and return of spontaneous circulation. Interestingly, in both of the studies that I just mentioned, there was not an association with a low end-title and mortality. So in pediatrics, I think we're all very, a little bit nervous to say that if your end-title is greater than 10, that you don't have any chance of a successful resuscitation. The other thing that you can use your end-title to do is actually pick up the return of spontaneous circulation. I presented this at SCCM a few years ago. There's several studies that have shown this. I like this one. It's an out-of-hospital cardiac arrest study of adults, 100, roughly, patients where they had pretty equal groups, 59 with a single episode of ROSC and 49 with no ROSC. And ROSC in this study was defined as having a palpable pulse. And this is actually looking at that time zero is where they actually had ROSC and the change in end-title. And you can see right around at time zero when they actually were able to feel a pulse, they actually had an abrupt rise in their end-title, and they quantified it in this study. It was about 10 millimeters of mercury. So end-title can actually give you a little bit of an idea from a flow standpoint of what's going on underneath your CPR. And the Hopkins group has been looking at this as actually being able to guide your actual resuscitation. And so they have a model where they actually look at feedback-optimized CPR. So they optimize kind of the depth using a marker, and they have a coach in some of their models. And they compare that to actually changing their CPR technique to improve the end-title during CPR. Their model, their primary model, is 20 minutes of asphyxia. They do induce these animals into ventricular fibrillation so they can study the CPR. They have a BLS period followed by an ALS period. But the real difference between is the adjustment of the technique to actually maximize end-title. And what they have been able to find is that when they do that and they target end-title during these resuscitations, the end-title group in the circles there, they've been able to improve the pressures, mean arterial pressure on the left, cerebral perfusion pressure on the right. And that ultimately, in this neonatal model, translated into a survival benefit for successful ROSC. They've now done more work actually looking at a pediatric model moving on from the neonatal model. Again, this is an asphyxial model. They use a target of 30 because the higher number was actually more associated with survival in their animals. And they actually gave us some data on how they're actually changing their CPR technique to try to get their end-title up. And so for the BLS period, what they'll actually do is they'll actually increase their chest compression rate by 10 every minute up to a max of 150. And what I liked here is once they hit the ALS portion, they will actually give epinephrine more frequently upwards of every two minutes to try to maintain that end-title greater than 30. And with this protocol, they have been able, again, the algorithmic or the end-title directed in the triangles here, they've been able to improve the end-title, improve, I shouldn't say improve, but they had higher chest compression rates in that group because that was the primary thing that they were actually changing. They were able to improve cerebral perfusion pressures. And then ultimately, when they looked across all the different levels of asphyxia that they did, they were able to see a large benefit in return of spontaneous circulation in this pediatric model. So there is some now pre-clinical data to show that end-title guided CO2 can improve outcomes. All right, so now moving into the ICU, and so now we have all the bells and whistles of ways that we may be able to monitor our CPR quality. And here's where I really think we can start thinking about how we could use blood pressure to guide our CPR. And so in the adults, a lot of this has to do with coronary perfusion pressure, and again, I'd like to go back to one of the classics by Norm Paradis, JAMA in 1990. This was a study of 100 pre-hospital and ED arrests where they looked at coronary perfusion pressure. And what they found is if the coronary perfusion pressure never got above 15, there were no survivors. And as your coronary perfusion pressure improved during CPR, you had higher rates of ROSC. In pediatrics, we've really focused on diastolic blood pressure as the more measurable and easily kind of obtained on our monitors type of target. This was a study done in the Capcorn Network by Bob Berg where we looked at 164 children who had an ICU arrest who had invasive arterial catheters in place at the time of the arrest. And we were really, we set out to look at really age-based cutoffs, these a priori cutoffs of 25 millimeters of mercury in infants and 30 in older children. What I'm showing you there on the left is a spline curve looking at mean diastolic pressure versus probability of survival. And as diastolic blood pressure increases, the probability of survival does. And the optimal blood pressure to discriminate survivors from non-survivors in ROC analyses was 34 millimeters of mercury. When we looked at our a priori cut points, we actually saw that survival was more than doubled when we hit these diastolic cut points. Fortunately, we've now been able to do an NHLBI-funded multicenter validation study published in Critical Care Medicine this year where we actually used those same targets in a much larger sample size, over 400 children. And ROC and survival to discharge, again, were significantly associated, or sorry, were significantly higher when these targets were met. Survival with favorable neurologic outcome just barely did not meet statistical significance, which after being at this conference, I'm thinking we probably need to use Bayesian analysis to look for probability because that seems to be what everybody does now. But clearly, these targets are associated with outcomes now in our validation study. And we've done some work in our lab at the Children's Hospital of Philadelphia actually targeting blood pressure during CPR to improve outcomes in animal models. And so in our model, similar to the Hopkins group, we have an asphyxial model where we clamp the endotracheal tube for seven minutes. These are anesthetized animals. At seven minutes, we unclamp the endotracheal tube and induce VF. And then we start studying our CPR groups. In the standard group, we do what the AHA tells us to do. Our depth is one-third the AP chest depth, or five centimeters in the adolescent adult model. But in the other group, we actually titrate the depth of compression to assist all the blood pressure. Depending on age, that target is 90 or 100. Couple minutes in, we start to administer epinephrine. We give it every four minutes in the standard care group. But in this HD CPR group, we only give drug if the coronary perfusion pressures are low. And we're able to give it more frequently than is currently recommended. So we can give epinephrine upwards of every minute and then give vasopressin rescue if two doses of epinephrine are unsuccessful at getting the coronary perfusion pressure up. We can kind of continue this process for the first 10 minutes, after which point we try to shock the animals out of VF. And we'll continue this same protocol for another 10 minutes before we end the experiment if we don't achieve ROSC. Everything else between the two groups is similar. The CPR rate, the ventilations, and the FIO2. And what we found in several models now, that if you actually target coronary perfusion pressure during CPR, over time you can improve it prior to the first shock and ultimately improve outcomes. In the solid line there, that's actually the CPP-directed group getting above that goal CPP of 20 millimeters of mercury by the end. And you can see in the guideline care group, kind of that lighter line, although we get a bump in the coronary perfusion pressure with the epinephrine doses, it quite never hits that 20 millimeters of mercury. And these improved pressures in that group ultimately corresponded to a huge survival benefit, 70% good neuro in the CPP-directed group versus lethality in our depth group. And there's been a lot of work by others, Ryan Morgan, Miriam Name, and others who have shown this across many models that this actually works. So it kind of makes you wonder when you're in the ICU, you may actually have both an end tidal and an ART line. So we actually took a look at this in 2016, where we looked at 60 of our animals where they had went through this protocol, and we used ROC curves to see which was the better discriminator between survivors and non-survivors. And not surprisingly, in our model where we're targeting blood pressure, diastolic blood pressure did perform better than end tidal CO2. And that was true across all of our models, both asphyxia and VF. And interestingly, in our animals, and this was work we had done first, the optimal target was 34 millimeters, which is exactly what we saw in the subsequent clinical study to discriminate survivors from non-survivors. So right here near the end, I want to talk a little bit about other non-invasive approaches to titrate CPR. So I think one of the things that probably we will be talking about in the future is pulse oximetry. So this was a study that was published in 2015 in PLOS One. I think a lot of us, when we do resuscitation, we get a pulse ox tracing, particularly in pediatrics. And we're really happy we have it, but we don't really know what it means when we're doing CPR. And so this is just what I'm showing you here is arterial line waveforms in the top panel and the corresponding pleth in the bottom panel. And not surprisingly, when you have ventricular fibrillation, there's no pulse, so there's no pulse ox. But then on the other two panels, you're looking at low-quality CPR versus high-quality CPR. And you can see that in low-quality CPR, you have a low-amplitude pulse ox. And when you have good blood pressures, you actually have better amplitude. And so they actually looked at this across the model, looking at high-quality and low-quality CPR. And not surprisingly, the physiologic markers, such as coronary perfusion pressure and end tidal that I talked about, were higher in the high-quality group. But then also simple things like pulse ox amplitude and area under the curve were also different between these groups. And so it is a universal monitor that probably in the future would have some benefit to be used as a CPR-quality monitor. And we've been leveraging a collaboration we have with the Villanova Center for Analytics of Dynamic Systems to look at pulse ox as a CPR-quality monitor. It was an ancillary to an R01. And after the information I just gave you, if you looked at those two tracings, I think everybody in the audience would say that the one on the left is the good quality compared to the one on the right. And what we've been doing is using machine learning to actually infer what the blood pressure is simultaneously recorded by just using the pulse ox simulator to see if our blood pressure is at goal, the similar targets that I talked about in the clinical study. And with the current work, which we're submitting to journal here very quickly, they've actually got an algorithm now. It's a little hard to see, but it's 90% accurate to infer whether or not the blood pressure would have met these targets based on just looking at pulse ox alone. All right. So in the last 41 seconds, if I had to pick just one thing to look at during CPR, I'm not going to just pick one thing. There probably isn't a single monitor. Like I said at the beginning, you need to use what you have, and a lot of this is going to depend on your location. And in 2023, if we're talking about one monitor, we've probably missed the boat. We really need to use a multimodal approach. As an example, just quickly, if you're doing CPR and your blood pressures are great, but your end title is five, we probably should be thinking about, did the tube come out? Is there a pulmonary embolus? Is there some kind of pathophysiology that we need to deal with? And so really kind of integrating all the information that we have available is really going to be the best way to go. But what you have to do is you have to have the courage to alter your approach. If things aren't working and you're not meeting the current physiologic target that you want, you really have to change up your technique. And so with that, I will leave you with the obligatory picture of my kids in our happy place, Cape May, New Jersey, which is quite the juxtaposition of this large conference nerve-wracking talking to you all. So thank you. Thank you.

CPR quality monitoring

rescuer-centric resuscitation

patient-centric resuscitation

CPR quality monitors

end-tidal CO2 monitoring

multimodal CPR monitoring