BreatheTV Episode 23
HFNC Clinical Considerations in the Pediatric ICU & Pediatric Emergency Department | A Webinar with Alexandre Rotta, MD, FCCM
Jeff Maglin: Hello everyone. My name is Jeff Maglin, I am the Director of Marketing for Vapotherm. Welcome to the webinar High Flow Nasal Cannula: Clinical Considerations in the Pediatric ICU and Pediatric Emergency Department led by Dr. Alexandre Rotta.
Before introducing our speaker I would like to go briefly over some housekeeping notes. If you would like to ask a question type in the question box in the webinar control panel window. After the presentation Dr. Rotta will answer as many questions as time permits. CEU certificates will be issued within one week pending receipt of evaluation. We have made improvements to our evaluation system and now you will be able to complete it electronically. If you are watching the webinar with several colleagues please forward the post-webinar email onto those colleagues so that they can complete their own evaluation form. Each individual must complete their own evaluation form in order to receive a certificate.
As a disclaimer, Vapotherm Hi-VNI Technology is tool for treating the signs and symptoms of patients that present in respiratory distress and for whom caregivers desire to add heat and moisture to breathing gasses. The views and ideas presented in this presentation are solely those of the speaker and individual results may vary. Vapotherm does not practice medicine or provide medical services or advice. Providers should refer to full indications for use, operating instructions, and/or prescribing information of any products referenced in this video before exercising their independent medical judgment to use or otherwise prescribe the products. Finally, Dr. Rotta is a paid consultant of Vapotherm.
It is now my pleasure to introduce our speaker Dr. Alexandre Rotta. Dr. Rotta is Chief of the division of pediatric critical care and the Linsalata Family Chair in pediatric critical care in emergency medicine at University Hospitals’ Rainbow Babies and Children’s Hospital. He also is Professor of pediatrics at Case Western Reserve University School of Medicine. Dr. Rotta is board certified in pediatrics, in pediatric critical care. His special interests include pediatric critical care, cardiac critical care, mechanical ventilation, and acute lung injury. He has authored or co-authored more than two dozen textbook chapters and nearly a 150 papers and abstracts in peer reviewed journals, more than a dozen on high-flow nasal cannula and bronchiolitis. He also is a frequently invited presenter at medical meetings and symposiums across the United States and internationally. He’s a reviewer for many of the leading medical journals in the field including Critical Care Medicine, Pediatric Critical Care Medicine, and the Journal of Critical Care. Dr. Rotta is a fellow of the Society of Critical Care Medicine and a member of the Pediatric Cardiac Intensive Care Society and the American Association for Respiratory Care. My great pleasure, and I will now hand the webinar over to Dr. Rotta.
Alexandre Rotta: Thank you, Jeff. I want to welcome our English-speaking audience in North America and throughout the world. We have a lot of ground to cover today. We’ll start with our goals and objectives.
Our goals today will be to understand the pediatric patient population who can benefit the most from high-flow nasal cannula support. We’ll understand some of the clinical data to support the use of high-flow nasal cannula in this pediatric patient population and talk about some therapy application guidelines.
Our objectives will involve understanding our evolving knowledge of how high-flow cannula support works. We’ll talk about how it is about the pressure but kind of it’s not, primary patient applications, clinical studies, and some comparative studies when we place high flow therapy support against traditional non-invasive ventilation such as CPAP and BiPAP as a first modality or as a support after extubation. And we’ll also share some data from a recently completed pilot randomized controlled trial of high-flow nasal cannula support versus CPAP and also some clues about how to start the therapy.
A high-flow nasal cannula system is predicated on four pillars, one of them being an “open” delivery system. What I mean by an open system is something that allows for leakage of gas around the cannula, so very much not something that looks like CPAP or BiPAP where you’re looking for an occlusion of gas leak and pressurization. In high-flow nasal cannula systems you want an open system to allow for gas to escape around the cannula. Gas needs to be optimally humidified and temperature controlled and I will refer to a gas that is at optimal humidity and temperature as conditioned gas throughout this presentation. A high-flow system is also predicated on high volumetric flows. Those are supraphysiologic and flows of dry gas that would, without the high-flow humidification, be quite desiccative to the airway. So these high volumetric flows have to be conditioned and, in case of the high-flow system by Vapotherm, high gas velocities which are achieved by pushing high volumetric flows through a paper cannula that allows for deep penetration into the airspace.
Regardless of which high-flow system is being utilized they all share a certain profile. You are using a water reservoir to create inflow to a humidifier that is heated to allow for optimal conditioning of the gas. And as gas that is conditioned leaves the machine the circuit that takes the gas to the patient needs to be temperature-controlled or otherwise the gas would cool down and there will be rainout. So the Vapotherm system does so in a very sophisticated and yet simple way by having all these elements, including the whole fiber humidification cartridge and the heating elements and pumps, inside of the precision flow device and uses temperature-conditioned water bath to insulate the circuit that takes the conditioned gas to the patient.
We’ll talk a little bit the mechanisms of action of high-flow cannula support. And this is the second dog analogy I will make today. This is a paper that was published in 1993 in The New England Journal of Medicine and it has nothing to do with a dog but I thought it would be abasement to put a patient with their head sticking out of the window. This tells a story of a 51-year-old woman who showed up in an emergency department in acute pulmonary edema bronchospasm respiratory failure with significant respiratory acidosis and hypoxemia. This patient was intubated and treated with conventional therapy. But when she recovered she was able to recount the story that while her husband was driving her to the hospital she was in such severe failure that she stuck her head out the window and every time that the car slowed down she became more symptomatic and as the car accelerated she had some relief. So physicians at that hospital were able to calculate that based on the speed of the vehicle at 80 miles an hour with her mouth open this individual was being exposed to a pressure of 7.8 centimeters of water. So driving a car down the highway at 80 miles an hour is equivalent to applying a CPAP of nearly eight to a patient in respiratory failure. So it is easy to try to conflate the concept so that perhaps this is how high-flow works, but it really isn’t because as you stick your head out the window like this pooch over here you notice that every orifice of that face is being exposed to that wall of gas coming at it with the same speed. In high flow therapy that doesn’t happen. And because of the leakage around the system there is very little pressurization that actually happens deep in the airway, and we’ll explore a little bit of that in more detail to follow.
To further complicate the understanding of the relationship between flow and pressure, there are several studies in the literature that show increasing pharyngeal pressures with increasing gas flow. This is a graphic from a study by Chris Milesi and his group, looking at the linear relationship between pharyngeal pressure and the gas flow through a high-flow cannula in liters per kilo per minute so that at about two liters per kilo per minute one would have a pressure of just over five centimeters of water in the airway. However, the key component here is pharyngeal pressure. This is a cross-section of a human’s face. And you can see here the turbulent flow generated by a high-flow system that penetrates deep into the airway. So if one places a probe in the retropharyngeal space that probe, that measures pressure, is being directly pressed on by the gas and that generates a pressure. But that doesn’t mean that this pressure that is leaking through the mouth and around the nostrils will actually pressurize the lower airway which is really what one is interested in. So yes, there is generation of pharyngeal pressure during high flow but the pressure that actually transmits to the lower airway and the subglottic space is negligible.
The same authors published this study showing in the top panel the pharyngeal pressure measurement during exhalation and inspiration. In the lower panel the esophageal pressure measurement which is a surrogate for intrathoracic pressure. As you can see on the left-sided panels, the pharyngeal pressure and esophageal pressure follow a same cycle with the esophageal pressure showing great negative inspiratory deflections showing that this patient is in significant distress. There is a negative inspiratory force of minus 40 centimeters of water with every breath and this pressure turns slightly positive upon exhalation. Upon initiation of high flow at 7 liters per minute there is a significant increase in pharyngeal pressure but no increase of esophageal pressure. So what that shows over here is that high flow was not a contributor to increasing intrathoracic pressure, but what it does is it decreases the negative pressure that occurs during each respiratory excursion. So it is obvious, by comparison of the lower panel, that the patient on the left has a great deal of respiratory distress and is using a lot of [inaudible 00:12:03] breathing whereas the patient on the right, on high flow, is showing attenuation of that dip in esophageal pressure which is really how high flow therapy works.
But that’s in patients and we wanted to further delineate that in an experimental model. So this is a study that was done by Rich Speicher and several colleagues here in Cleveland looking at an experimental model of airway obstruction, subjected to high flow therapy. To that effect, we used a 3D-printed head with a tracheal attachment that was interposed to various resistors: a low resistor, a medium resistor, and a high resistor. And then this head was ventilated with a high-flow cannula and a volunteer was breathing through this tube while having an esophageal monitor attached. And the experimental conditions involved no flow, flows of 10, 20, and 40 liters per minute with and without Heliox.
We performed a total of five experiments for each experimental condition and we used the change in esophageal pressure and respiratory rate to calculate the pressure rate product. A pressure rate product is the multiplication of respiratory rate and esophageal pressures and it’s very good marker of work of breathing.
If I take your attention to the right of this graph, you will see that under high resistance the pressure rate product here in blue is very high. This means that that volunteer was having great difficulty with tachypnea and high respiratory effort. And this condition, when we scored the [inaudible 00:13:54] scores for this volunteer, it was a [inaudible 00:13:57] score of 10, meaning that this was an unsustainable way to breathe for more than a minute. This simple application of high flow in a very controlled scenario, nearly half the pressure rate product, so it made it a lot easier to breathe. And this was a 10 liters per minute in red, 20 liters per minute in green, and 40 liters per minute in yellow, showing a dose response alleviation of respiratory distress when these volunteers were subjected to high flow therapy. When we added Heliox to the mixture, again, coming from a very high baseline pressure rate product, the addition of high flow at 10, 20, or 40 liters per minute showed significant attenuation of respiratory effort.
This is to exemplify coming from baseline with and without Heliox – without Heliox in blue, with Heliox in red. You come from a very high pressure rate product. This simple application of high flow in blue causes a significant reduction in work of breathing. And if we use, in this model of airway obstruction, Heliox as the gas mixture there is a significant attenuation again in comparison to just oxygen.
The pressure rate product has been used to also clarify mechanistic approaches to high flow utilization in infants. And this is our data from colleagues at the Children’s Hospital Los Angeles looking at esophageal pressure at end-exhalation. So this is the pressure that is transmitted in equilibration to the alveoli. And showing that while the patient is intubated with a PEEP of around five or six, the baseline esophageal pressure was around five or six. Those patients were then extubated to simple cannula and then started on a high flow from 2, 5, and 8 liters. And you can see that the accrual of pressure from breathing a standard cannula at 4.5 centimeters of water to high flow at 5.4 centimeters of water was less than one centimeter of water. So the pressure contribution of high flow in the intrathoracic cavity is very low and yet, by incrementing flow, they saw a significant reduction in the pressure rate product or the work of breathing of this patient is very much like we saw in the adult healthy volunteers in the previous model.
The primary application for high flow therapy in pediatrics is the patient with bronchiolitis. These patients are abound, they permeate our units through entire respiratory season, and they show benefit from being treated with high flow therapy.
Bronchiolitis is an acute inflammation of the bronchioles and it’s characterized by [inaudible 00:16:57] bronchospasm, by mucosal swelling and mucus buildup that causes an airway obstruction. It is typically caused by a viral infection such a respiratory syncytial virus and is very prevalent in young infants. In fact, it is the leading cause of hospitalization for infants in North America.
Most cases of bronchiolitis result from a viral pathogen with RSV being the most prevalent one, but also parainfluenza, influenza, adenovirus, and human metapneumovirus are frequent offenders.
RSV bronchiolitis is highly contagious. It is transmitted by direct contact with nasal secretions from subject to subject or through fomites. It is viable for 30 minutes on skin and approximately six hours on fomites. So it’s a highly contagious disease. In fact, each year nearly a 125,000 children are hospitalized in the United States with RSV.
It is a seasonal disease. About 93% of these cases occur between November and April in North America, and reinfection is quite common. Florida, because of its poorly defined winter, has a longer season spanning from late August all the way to April.
The epidemiology of bronchiolitis is quite astounding. If you take children under six months of age, out of a 1, 000 children infected with RSV there will be 55 emergency department visits and 17 hospitalizations. Most pre-school children will be infected by RSV in one of those respiratory seasons and approximately one third of children will develop bronchiolitis in the first two years of life. In spite of its very high prevalence, bronchiolitis has a very low mortality. So if you think of nearly of a 125,000 cases a year in North America, there are less than 400 annual deaths.
So one could ask, “So what is the big deal? You have a patient that then develops acute respiratory failure from bronchiolitis and they will eventually survive.” What does matter is the path you can get there. You might have a patient that just has a simple oxygen cannula and goes up that hill and survives. You might have a patient that is affected by acute respiratory failure, requires high-flow cannula support and survives. Or you can have a patient that requires mechanical ventilation and has a much more convoluted course but also survives. Well trust me, this is quite important. The path you take actually means a lot.
Steve Shein, one of my colleagues here in Cleveland, has done a lot of work in trying to characterize what happens to patients with bronchiolitis as it relates to their exposure to neuropharmacological drugs and their neurological morbidity. This is a study that was recently published in Pediatric Critical Care Medicine looking at the number of drugs that a patient with bronchiolitis is exposed to over the course of the years, from 2006 to 2015, using a very large sample. It is remarkable to note on the right panel that in 2015 more than 55% of patients with bronchiolitis admitted to an ICU have five or more drugs or drug classes for sedation and analgesia employed during that hospital stay.
And this is not something that has no human cost. A follow-up study, also by Steve Shein, looking at over 3,700 PICU patients through the Virtual PICU database and nearly 10,000 ICU patients in the PHIS database tried to characterize the impact of bronchiolitis and the bronchiolitis support on neurological and functional morbidity. This was a study that was, again, quite large, involving nearly 14,000 patients altogether. These were high-risk patients mostly in the two to three month age group category. They were mostly male, about 60% of them. As it is traditional in clinical studies with bronchiolitis, males tend to be more severely affected. And mechanical ventilation rates were somewhere between 15% to 21% depending on the sample, and the length of stay for the hospital was four days and for the ICU was two days.
What they showed was that children who came in healthy to the hospital with bronchiolitis as the only factor for hospital admission, if they required intubation 27% of them had a neurofunctional morbidity according the Virtual PICU database or 14% of them according to the PHIS database, which was significantly higher than if the patient did not require mechanical ventilation. So what this study shows is that avoidance of mechanical ventilation and sedation and analgesia in these very vulnerable bronchiolitic patients should be tried at all cost and high flow therapy might be a path to that.
I’m going to use some clinical studies to just, again, characterize where we are with the treatment of bronchiolitis and whether or not high flow has made a different.
This is a study by Bryan McKee, one of our former fellows here in Cleveland, looking at trends in treatment and outcomes for bronchiolitis in the ICU. So the goal was to really try to characterize what we do in patients with bronchiolitis in the critical environment, and for that we used a large sample spanning from 2009 to 2015 using the Virtual PICU Systems which is something that encompasses over a 120 critical care units throughout North America and the world, amassed about over 38,000 patients with critical bronchiolitis, aged less than 2 years. Again, most of those patients were male, nearly 60%. The majority of them were previously healthy, 58%. Some of them had co-morbidities: cardiovascular, pulmonary, or prematurity. Mortality, as expected, was very low, 0.63%. The majority of patients that came to an ICU with critical bronchiolitis were supported by high-flow cannula support – 44%, 23% required mechanical ventilation, just over 1% required high-frequency ventilation, and a negligible minority that required extracorporeal life support.
As I mentioned before, the probability of death with critical bronchiolitis is very low. The pediatric index of mortality probability of death is around 0.25% and the PRISM probability of death or pediatric risk of mortality is about 0.5%. And that remained constant throughout the study period and so has the effective mortality. There are some ups and downs but it’s largely less than 0.75% throughout the years.
What is interesting in the approach to the treatment of bronchiolitis is that when you compare 2008 to 2015 there is a significant reduction in the utilization of mechanical ventilation from about 25% to about 19.9% nationally, with no increase in utilization of non-invasive ventilation but a significant increase in the utilization of high-flow cannula support. So as high flow use has increased the need for mechanical ventilation has decreased, and this is in a contemporary sample.
These data are not in isolation, there is now plenty of clinical evidence to support this same phenomenon. This is a paper from 2010 Journal of Pediatrics looking at patients under two years of age with bronchiolitis observed in a cohort in Massachusetts that was prior to and after high-flow cannula support availability. So in the cohort on the left, 2005-2006, this is before high flow was available at this institution so no patients received high flow. In the 2006-2007 cohort 87% of patients received high flow. When you compare these two time periods, the earlier time period patients had a median age of three months versus two, had an admission respiratory rate of 47 versus 61, and had RSV positivity just over 50% versus 69%. So one could argue that the high flow epoch was no different and possibly sicker based on smaller weights, high respiratory rates, and higher positivity for RSV than the controlled sample.
When you look at the pre-high flow era, those patients required intubation in 23% of the cases versus 9% when high flow was utilized.
To further confirm these data, these are data from Australia and New Zealand looking at nearly 300 children admitted to the ICU under two years of age, the majority of whom had bronchiolitis. And this describes the use of high flow and need for mechanical ventilation in this cohort. What you see here from 2005 to 2009 in blue is the percentage of patients coming in to that intensive care unit with critical bronchiolitis who required high flow. And as you can see, from 2005 to 2009 there has been a steady increase in the use of high flow in the institutions in that multi-center study. And at the same time as high flow use went up the need for mechanical ventilation steadily went down to the point where from 2005 37% of critical bronchiolitis patients required mechanical ventilation versus 7% in 2009.
Now, one can argue that perhaps time played a role here, that perhaps the season in 2005 was not as severe as the season in 2009 or that we simply got better at taking care of patients with bronchiolitis between then and now. To offset that criticism the authors brought in a cohort of units from the same geographical region who did not use high-flow cannula support in the same contemporaneous period. So in 2009, for units that did not use high-flow cannula the need for intubation rate was still 28%, which lower than 37. So this is the effect of time, this is the effect of high flow.
High-flow therapy is one of those modalities that show its response fairly early. These are data from Schibler and colleagues, looking at heart rate and respiratory rate prior to and upon initiation of high flow. The top graph or heart rate shows an arrow which is when high flow was started and follows responders and non-responders for the next several hours. You can see that those patients who will clinically respond to high flow within 30 to 60 minutes have a significant decrease in their heart rate, which is persistent throughout the observation period. And the same is true for respiratory rate, those responders will have a decreasing respiratory rate from baseline that will be sustained throughout the observation period. Those patients who do not respond and require escalation support continue to be tachycardic and tachypneic and will eventually require additional support in the form of non-invasive ventilation or endotracheal intubation.
Our group in Cleveland has been busy trying to understand what happens when you start someone on high flow versus non-invasive ventilation for a primary patient that presents with critical bronchiolitis or with those who have been intubated because of respiratory failure.
These are data on all comers with bronchiolitis to an intensive care unit, again, using the Virtual PICU database, using more than a 125 intensive care units worldwide, mostly in North America. We chose patients with primary diagnose of bronchiolitis with ages between and two years of age and we picked up all patients who received high-flow cannula or non-invasive positive pressure in the form of CPAP or BiPAP. Our primary outcome was the need for intubation, which is something that is unequivocal – you either need a breathing tube or you don’t. And the second outcome was mortality and ICU length of stay.
We had 5,562 patients in the high flow group and 934 patients in the non-invasive ventilation group. The need for intubation in high-flow nasal cannula patients was significantly lower than those in the non-invasive ventilation group, 11% versus 20%. Mortality was also lower in high flow patients and length of stay was also lower, 2.8 days versus 3.74 days.
Now, we understand that the choice of non-invasive ventilation as a primary modality might be dependent on associated factors so we performed a bivariate analysis looking at various clinical factors to try to understand what is the difference between who were originally started on high-flow cannula versus those on non-invasive ventilation. It turns out patients that were started on non-invasive ventilation were younger, 3.2 months versus 4.1, they were lighter, there was some racial differences within those groups, and also there was a slightly higher predicted pediatric index of mortality risk for the non-invasive ventilation group.
So we felt that we needed to correct for all those and we performed then an adjusted odds ratio correcting for age, weight, race, and risk of mortality or any comorbidity. And we still found that the primary application of high-flow cannula decreased the need for intubation significantly such that by starting a patient on non-invasive ventilation as a first modality accrued an odd ratio of 1.5 greater chance of intubation than if you used high flow. The same is true for mortality, and the length of stay differences disappeared once we corrected for the multi-variable correction.
Another way to approach the views of high flow is on patients who came in intubated or required having endotracheal intubation and are being weaned from mechanical support and can then be assigned to either high flow support or CPAP/BiPAP support upon extubation. So Salar Badruddin, one of our critical care fellows, and our research group conducted this study with the objective of trying to understand what are the clinical outcomes of patients who received high flow support or non-invasive ventilation following extubation. They, again, used the Virtual PICU System database, included patients under two years of age with critical bronchiolitis who received high flow or non-invasive ventilation following extubation. So we didn’t include any patients who were just extubated to room air or a simple cannula. The primary outcome was the need for re-intubation within 24 hours of extubation, and the secondary outcome was PICU length of stay.
When we look at differences between patients who were re-intubated versus non-intubated, first of all, when you take a large sample like that the failure rate was approximately 6%, which is very much in line with what’s expected from patients recovering from acute respiratory failure from bronchiolitis. There were no differences in age, weight, pediatric index of mortality, gender, racial, or comorbidities, and the only difference between patients who required re-intubation and those who did not was that those who required re-intubation had a longer PICU length of stay.
If you take the entire cohort of patients who were extubated, the vast majority was extubated to high-flow cannula support, 656 patients, and a minority at a 103 patients received non-invasive ventilation. The likelihood of being re- intubated was significantly lower for patients that were extubated to high-flow cannula support, 5% versus 12% on the non-invasive group.
And we understand there might have been confounders here so we conducted a multi-variable analysis looking at all comorbidities and using the PIM2 score as calibration for severity of illness in this cohort. And the only remarkable significance factor of extubation failure was whether or not a patient was extubated to high flow, showing that high-flow cannula support significantly decreased the likelihood of intubation even when all other comorbidities and risk of mortality variables were entered into the model. So in this case high-flow cannula support was protective of the need for re-intubation upon extubation.
I want to spend some time talking about this study. We were very excited when the study was conducted and eventually published. This is the TRAMONTANE study, again by Christophe Milesi and the group in France. It’s a multi-center study looking at patients with acute viral bronchiolitis and it was published in Intensive Care Medicine last year. This is a very well-conducted study, and because of the weight of the journal and the expertise of the folks who conducted it, it should really get your attention.
But if you are the type of person who reads the title of a study or the first figure of a study you might have the wrong impression from this study. This is figure one of the study showing the probability of failure of patients who received high-flow nasal cannula support or nasal CPAP, those were the two comparison groups. And as you can see here in red, high-flow cannula support had a higher percentage of failure than non-invasive support like CPAP, 50% versus 30%. So if the only thing you look at on this paper is this figure you would surmise that high-flow nasal cannula is inferior to nasal CPAP in the support of patients with acute viral bronchiolitis.
But really, the devil is in the details here and you need to dissect this file a little bit more. First of all, what was failure for this study? Well failure for this study is not how I think of failure, which is the need to escalate support or the need to intubate the patient. Failure here was a composite end point made by several scores, some of them are clinical, some are more objective, some less objective, and it was the aggregate of those scores that decided whether or not a patient has failed.
If we look at the patient distribution in this study, a 142 patients were included, 71 were randomized to the nasal CPAP group and 71 were randomized to the high flow group. Those patients in the high flow group, 36 of them failed and were switched to CPAP and 35 of them succeeded. Those who were switched to CPAP – so this was a crossover study – CPAP was able to rescue 72% of them whereas 28% of those patients failed and required intubation or BiPAP or higher flows.
Now, let’s focus on the nasal CPAP group. The nasal CPAP group had 22 patients who failed, so less patients failed compared to high flow and that’s what that figure that I showed in the slide that preceded this one showed, However, if we look at those patients who failed CPAP and were switched to high flow, high flow was able to successfully rescue 82% of those patients. So how can we say that high flow is inferior to nasal CPAP when it is rescuing the vast majority of patients that failed the supposedly preferred strategy?
And when we look at actually what I would consider failure, which is the need to escalate to CPAP, BiPAP, or intubation, there is really no difference between the groups that were originally assigned to CPAP and the groups that were originally assigned to high-flow cannula, so that in these 142 patients only two had to be intubated and only 10 had to be escalated to BiPAP. So what I think that shows is that CPAP and high-flow cannula are very equivalent challenges that some patients will be assigned to nasal CPAP and fail and they can still be rescued by high flow therapy, and some patients will be assigned to high-flow and will fail and might still be rescuable by application of CPAP. And if that fails you still have the option of using BiPAP, with intubation being a very far-reaching modality for this patient population.
We attempted to answer some of these questions in a setting where there was clinical equipoise to do so. We attempted to run a randomized trial of high-flow and CPAP in our unit and we simply did not have clinical equipoise to withhold high flow therapy in patients who came in with critical bronchiolitis. But our colleagues in Brazil were facing the opposite dilemma. They had CPAP as the most ingrained strategy and they were hesitant to use high flow, so we thought there was equipoise there to randomize patients to high flow therapy versus CPAP. So Dr. Cesar and Dr. Horigoshi and colleagues at Hospital Infantil Sabara in Sao Paulo conducted a beautiful pilot study showing the feasibility of doing so in a developing country with a team of critical specialists that are very invested in the care of their patients, so support collaborators.
The original hypothesis was that high-flow cannula support will be equivalent to CPAP in high-risk infants with critical bronchiolitis. They enrolled infants, again high-risk, under nine months of age with critical bronchiolitis in this very well-established tertiary PICU in Sao Paulo, Brazil, in the 2016-2017 respiratory season. Patients were randomized to receive high-flow cannula of up eight to liters per minute and an FiO2 that was titrated to achieve a saturation greater than 92%, or randomized to receive CPAP six liters per minute, again with FiO2 titrated to normoxia.
The primary endpoint was treatment failure or escalation to BiPAP or intubation. So, the treatment failure was not a composite score, it was a hard end for need to escalate to BiPAP or intubation. And the secondary endpoint were duration of support, PICU length of stay, and hospital length of stay. They enrolled 62 patients, 28 in the CPAP group and 34 in the high flow group.
The age, weight, and RSV and severity score was similar between the two groups. Again, this was a high-risk cohort, ages between 2.4 and 3.4 months, about 5.5, 6 kilos, very high percentage of RSV positivity. So again, a sick cohort. And they came in with high bronchiolitis severity scores.
The treatment failure rate, again, defined by need to escalate to BiPAP or intubation, was similar to between the two groups, nearly 36% in the CPAP group and nearly 38% in the high flow group, a difference that was not statistical difference. The duration of support was three days in the CPAP group and three days in the high flow group, so two days and three days. The PICU length of stay was identical between the two groups, five days. And the hospital length of stay, eight to nine days, again not statistically significantly different. So showing that in that environment randomization of patients to receive high flow had the same clinical outcomes of those patients who were randomized to receive CPAP as a primary support modality. So how does high-flow nasal cannula therapy impact bronchiolitis? Well for one, the conditioned gas at proper temperature and humidity allows for delivery of very high flow rates that would not be possible if gas was not conditioned. If you attempted to do so without heating and humidifying gas and give 30 liters per minute to a teenager, within an hour there will be desiccation of the nasal mucosa and airway epithelium because you just cannot transfer that much heat and humidity to such high flow rate. So the conditioned gas allows for delivery of very high flow rates.
That delivery of high flow rates into the nasal cavity bypasses the highest resistor of the upper airway which is the nasal turbinates and nasal cavity. So, you’re essentially channeling flow past the nasal cavity into the nasopharynx and oropharynx almost as if it were a conveyor belt, showing that the work of bypassing that resistor is borne by the cannula.
By introducing such supraphysiologic flows into the anatomical dead space you are essentially flushing it from carbon dioxide. So you are decreasing anatomical dead space and re-breathing your CO2 that would happen during normal breathing.
The application of high flow does provide some level of pharyngeal pressure but very low levels of intrathoracic pressure, but there is some pressure.
The bottom line is high-flow cannula support decreases your energy expenditure. It reduces work of breathing by decreasing your pressure rate product, so patients don’t have to breathe as fast or as hard as under normal conditions. And it also decreases the work involved in heating and humidifying gas. That humidity and perfect conditioning of gas improves mucociliary function which has implications in the mucociliary elevator and mucus transport.
It reduces the need for mechanical ventilation and it reduces the exposure to sedatives and analgesics that come with a child requiring mechanical ventilation. So for all those reasons high-flow nasal cannula support has a firm place in our therapeutic arsenal in North America.
How to start high flow is something that evolved over the past few years. I think data from CHLA and the French group has informed us on how to best match the need for support to certain flows. And it is visible through this table that as patients get older, from the neonatal period to infants to young children to school-aged children to teenagers and adults, there is an increase in weight and an increase in flow. So we will typically start a newborn at about 4, 5 liters per minute and we will peak at about 8 liters. The patient up to 10 kilograms, we will start them between 4 and 10 liters per minute, we’ll peak at 2 per kilo per minute which is 20 liters. That curve starts to flatten a little bit so that a child that is one to six years of age will generally start somewhere between 5 and 16 liters and will go as high as 30. Patients between six and twelve years of age, 20 to 40 kilos, will generally start somewhere between 10 and 20 liters, and we can go as high as the maximum flow of 40 liters per minute. And the same is true for the teenager and adult-sized patients where we’ll typically start around 20 to 30 liters per minute, going as high as 40. So to summarize what we discussed today, high flow use has significantly increased over the past few years, and concomitantly the need for invasive mechanical ventilation has decreased.
The primary pediatric application of high flow is bronchiolitis, and that is heavily supported by literature and new papers come out just about every month supporting its use in bronchiolitis. There are several other applications that we did not discuss here today, but the use of high flow has really permeated ICUs throughout North America and the world.
Heliox delivered through high-flow cannula has a place in the treatment plan, particularly in those patients with airway obstruction.
High flow therapy compares favorably to CPAP and BiPAP as a first support modality and also as a backup modality for patients who were separating from mechanical ventilation after extubation.
A pilot randomized clinical trial of high-flow cannula support versus CPAP has shown promising results.
It is important to use the physiologic data that has been given to us by both the LA group and the French group to help us set the proper flow to maximize effect.
We’ve been talking for approximately 50 minutes now and I think we will have time for questions. I hope some of you have been able to enter questions within our question box, and we’ll entertain some of them.
Jeff Maglin: Before get to our questions I just want to make one reminder to all attendees to look for the email with links to the CEU evaluation forms to complete, and we’ll process the forms and generate certificates within a week. So those who do have questions, we have the question area of the monitor, please type them in and we’ll have Dr. Rotta try to get to as many as we can.
Alexandre Rotta: So I’m looking at a question here, “So in the graph showing up to five drugs, was oxygen or helium considered as a drug?”
No. So those are five sedative or analgesic drugs. I mean those patients were receiving many, many more drugs including oxygen. We don’t consider helium a drug, helium is inert so it doesn’t have a biological effect other than its physical effect. But that graph is astounding because those were children with bronchiolitis in the ICU and more than half of them were exposed to more than five drugs, and we’re talking morphine, fentanyl, [inaudible 00:50:54], Versed, Ativan, chlorhydrate, Propofol. So these are just neuropharmacological drugs and that’s why they caught our attention so heavily.
Another question here, “In your practice how do you determine when to use high-flow nasal cannula in bronchiolitis in the emergency department versus other respiratory modalities like standard option therapy, CPAP, non-invasive ventilation?”
So as it applies to emergency department proper, patients will come in with bronchiolitis and they will be assessed by triage and categorized into severity. Part of that assessment involves an assessment of heart rate, respiratory rate, and oximetry. So if a patient has significant distress as in marked retractions and impending respiratory failure this patients will be expedited into a room and will be assessed by a physician and a respiratory therapist, and a decision then will be made whether or not this patient requires simple oxygen to treat hypoxemia or if the patient requires additional support.
High flow therapy is our primary support modality both in the emergency department and in the intensive care units so we will use that as the first modality mostly because it is so easy to use, it’s easy to set up, and it doesn’t require a very complicated interface. For instance, an adult emergency department might be ambivalent as to using high flow versus a CPAP/BiPAP mask because an adult patient can be quite cooperative with non-invasive techniques. But those of you who practice in the pediatric emergency department know that finding the right match, the right patient interface with the non-invasive device through a pipe-feeding mask can be quite challenging, and high flow therapy eliminates that guesswork. So we generally start our patients with bronchiolitis in a respiratory failure requiring additional support on high flow. We have a high flow transport system that will then be capable of transporting those patients who responded to high flow, and that’s the majority of them, to the critical care environment where we will then switch them to a standard high-flow cannula, returning the transport system back to the emergency department for the next patient.
Another question is, “In your practice are there specific situations where you would place a patient with bronchiolitis on CPAP without considering high-flow nasal cannula? What about non-invasive ventilation?”
So we consider high flow therapy as being equivalent to CPAP so in our practice we don’t use CPAP as a primary modality on those patients. The type of patient that we will generally not use high flow therapy is the small infant, usually with some degree of prematurity, who comes in with bronchiolitis and has apnea from RSV, usually, as the etiology. Those patients are not good candidates for high flow and, fortunately, those are a minority of patients. But a patient who is apneic will require some level of support, so for those patients we will attempt some nasal or nasofacial bypass to assist breathing and, to be perfectly honest, to create some level of annoyance to that patient that will allow them to continue to breath while we either load them with caffeine or move onto the next therapy which will be then into endotracheal intubation. But the vast majority of patients we encounter with a respiratory failure, with acute critical bronchiolitis, will be places on high-flow cannula if they have the ability to breath.
We have a question about weaning. “What is your facility’s weaning approach to high-flow cannula in the patient with bronchiolitis and at what point do you discontinue use?”
So this has been a topic of great debate and there are largely three schools of thought on the weaning. One school is that you don’t really need a protocol, that you can just keep assessing your patient and frequent assessment will allow you to then decide when the patient is ready to wean off high-flow cannula. This is how we started. And then we realized that there is a lot of stylistic differences, that Dr. A and respiratory therapist B prefer to go this way whereas someone else is a little more aggressive and someone else is more conservative.
So in order to be able to study bronchiolitis and support modalities we decided to standardize our escalation and our weaning protocols in the ICU and actually create fairly rigid protocols that once a patient has been stable for a certain number of hours and is ready to start de-escalation, the patients get plugged into a protocol that is nurse and respiratory therapy-driven. These patients are assessed clinically and once an hour they are scored on a certain bronchiolitis score that we use, and if they meet certain criteria flow and oxygen are weaned until the people on low flow enough to be separated from high-flow cannula. How much or how low is that flow depends on patient size, but several hospitals have created various protocols for that and we developed our own and we use it, and we’re in the process of showing data that will be presented in an upcoming meeting that the utilization of a weaning protocol significantly shortened the number of hours of a weaning process of patients with bronchiolitis on high flow.
And then there is a different approach, and I’ll call that the Atlanta approach, where in their group they think that high flow a support modality that can be an all or nothing. So if a patient is ready to start weaning they will just go ahead and turn it off and see how the patient will do, and if the patient does well – great, and if the patient still needs support then they will go back. So this is more ripping it off like a band-aid, knowing that they can always fall back on high flow. Now, that group has published their data and they’ve been quite successful with this approach. This is not the approach we take. But what is pretty clear to us is that however you’re going to use high flow you should have a system of how you escalate therapy and how you separate from therapy as to remove variability and allow you to compare yourself with a different time point in your career or with other institutions.
There’s a question asked, “Do you place any high flow patients on the general [inaudible 00:58:27] floor or do they go to the PICU?” So I’ll tell you what we do and I’ll tell you what some other places do. At Rainbow Babies and Children’s Hospital the institution has made a decision that it is unacceptable for a patient have significant clinical deterioration while in one of the general floors. So for that matter, high flow therapy is only used in an ICU environment or in a step down unit environment. And the step down unit environment is a cardiac step down for patients with cardiovascular disease or recovering from heart surgery after they’ve been discharged from the ICU. So those are the only two places where we do high flow.
Now, it has nothing to do with how complex high flow is or with the need for assessment of those patients, the real reason for us is that if we think a patient requires high flow this is a patient that requires additional monitoring and attention because things might not go the way we want them, so we bring those patients into a highly-monitored environment.
Now, this is an institutional preference. I can tell you that there is nothing specific about high flow that makes it exotic. High flow is so simple to use it can be used just about anywhere: the emergency department, the floors, the PICU, and there are institutions that have done this very successfully and have published their results of using high flow up on the floors and as long as you have a system to frequently assess those patients everything goes well. There has also been studies, mostly out of Australia, looking at using high flow in remote places and keeping those patients at those hospitals without having to transfer them to a referral center which might be hundreds of miles away. And so there is great precedent of using high flow in non-critical environments, but at my institution we choose to do so.
I think we are past our top of the hour time and so I’m going to close here and will look at some of these questions, and if there is a way of answering them back to you I will do so. I think you for spending the past hour with me and I hope you have found our discussion helpful. Thank you very much.