and thanks the conference organizer for inviting us to present the early high flow in children with acute hypoxemic respiratory failure. This study is a study that comes out of a large body of work by the same group, particularly the PREDICT group in Australia and New Zealand. These are my disclosures, nothing that interferes with the presentation of the results. We studied in the past nasal high flow and bronchiolitis and we're still trying to find its way how to use it. We've got some evidence, particularly during the COVID-19 pandemic, and high flow has been sort of a standard therapy in NICU and PICU. But we don't have any good evidence in the older children than one year of age, and since children are not little adults, we need to generate our own evidence in the pediatric field. So we first performed quite a large randomized controlled pilot study to inform our larger randomized controlled trial. We performed this pilot data in two tertiary children's hospitals in Australia. They were aged between 0 and 16 and they all needed oxygen therapy because they were hypoxemic. As you can see, the majority of the children were in the age bracket 1 to 5, so 80% of these patients presenting to the emergency department are in this age bracket. Most of these children, roughly 70-75%, present with a WHEEZE, so obstructive airway disease with increased work of breathing. In the pilot trial, what we could show is that a treatment failure on their allocated oxygen therapy was slightly higher in the standard oxygen group and we had, interestingly, a little bit more ICU admission in the high flow group compared to the standard oxygen in the pilot trial. You can see only one child got intubated out of more than 500 patients. So that led us to the question what is a good candidate for a bigger trial as the primary outcome. Escalation of care is probably not necessarily the best outcome because parents do not mind to be admitted with their child to ICU, so that kicks out as well as ICU admissions. Intubation and invasive ventilation would be very interesting, but a very low incidence rate of intubations in the pilot, so we dealt then with length of stay in hospital as the primary outcome. The target population was then children between 1 and 5 because that's the majority of these patients presenting either with a WHEEZE or absent WHEEZE. They needed to be hypoxemic and, of course, they needed to be admitted to hospital. The trial started with oxygen therapy, obviously, in the emergency department and continued in the general pediatric ward. This is the flow diagram. You can see we had a large screen population and we enrolled then, in the end, more than 1,500 patients into either nasal high flow or standard oxygen therapy. Now, if you can go down in the flow diagram, you will see that from the nasal high flow group, a lot of children turned back to standard oxygen. More than 300 patients were flipped back to standard oxygen. I'll get back to this figure later on because it's a very important outcome. The trial was performed according to standards with a one-to-one random allocation to standard oxygen or nasal high flow in 14 metropolitan and tertiary hospitals in Australia and New Zealand. We could complete the trial just before the pandemic hit us because that would have jeopardized the study flow and, as well, the use of nasal high flow. As I said, primary outcome was length of stay. I won't bore you too much with the baseline characteristics. They are very similar in both arms and very well distributed. That just leads to a good study performance. Now, the primary outcome. Nothing was significant in comparison to standard and nasal high flow. I highlight here the length of stay in nasal high flow was 1.77 days in standard oxygen 1.50 days. That makes a six and a half hour difference in the length of stay. We wanted to demonstrate that nasal high flow reduces the length of stay in hospital, which we couldn't demonstrate. The rest, as well, of the primary outcome allocated or investigated in the different groups, like absent wheeze or present wheeze, there's no difference at all between the two groups or, in other words, nasal high flow wasn't superior to standard oxygen therapy. Now, if you get to the secondary outcome, I just highlight a few of them that are important to understand and may change your practice in future. First of all, if you look at the ICU admissions, there was a significant greater proportion of patients in the nasal high flow arm that were admitted to intensive care, which was, to us, again, very surprising. If you look at do they need more invasive ventilation or non-invasive ventilation, there's no difference in the two arms. If you look at four patients out of more than 1,500 patients needed invasive ventilation, so sort of the studies performing well, we couldn't demonstrate a difference in the two groups, but, however, we can say that we have a very, very low intubation rate in this population. Now, another aspect is that some of these hospitals didn't have an on-site ICU. So, for a child that needed ICU treatment, these children were transferred in an ambulance, rotary, or fixed wing to the next tertiary children's hospital. So, if we look at the transfer rates to an EPQ, you can see that we have only 2.3% in the nasal high flow and 2.2% in the standard oxygen arm who needed a transfer. So, much lower than the overall population, and the overall population was driven by hospitals who have direct access to intensive care. And you can see as well in this subgroup of hospitals without on-site ICU, the length of stay was shorter in the nasal high flow arm compared to the standard oxygen arm. I'm not giving any p-value because that wasn't pre-planned in the analysis. Now, a second look at these PICU admissions. You can see first the total cohort, 12.5% in the high flow arm, and 6.9% in the standard oxygen arm. In the hospitals who have an on-site ICU, this number climbs up to 13.8% compared to 7.6% in the standard oxygen arm, whereas hospitals who have not a direct access to an ICU have a much lower, so almost four-fold lower, ICU admission compared to the tertiary children's hospitals. Now, I highlighted in the consort statement a flow diagram that a lot of children got switched back from nasal high flow to standard oxygen, and you can see that here very clearly, that tolerance was a problem. So, a lot of these children didn't tolerate or didn't like the nasal high flow, which is very different to what we observed in the previous trials. Now, the reason why and how they got changed back to nasal high flow, you can see when a change in oxygen therapy occurred in the standard oxygen arm, it was because they got sicker. They had an increased work of breathing, where in the nasal high flow arm, the kids were actually better and didn't tolerate any more the nasal high flow, so they ripped it off because they don't like it. Now, how does this trial compare to our previous trial in bronchiolitis? So, the previous trial was PARIS-1, was performed in infants less than one year of age with bronchiolitis, and they were as well hypoxemic. We performed this trial as well in 17 metropolitan tertiary hospitals. However, the outcome was different. We looked at escalation of care, but if we look at the secondary outcomes in the PARIS-1 trial compared to PARIS-2, we can see as well that the duration, the length of stay was slightly longer in the high flow arm with 3.12 days compared to 2.94 days, and the difference is 4.3 hours. So again, when a patient is placed on nasal high flow, it seems that they stay just slightly longer, but it can be relevant if you look at patient flow within a hospital. And here again, the ICU admissions, we have more ICU admission in the PARIS-1 trial compared, if we compare nasal high flow to standard oxygen. And if you compare here again, the transfer to ICU from another hospital in both arms, the ICU admission proportion is exactly the same. So, which means that obviously these peripheral hospital outperform the tertiary hospitals because there's no difference in severity of the disease. So, in conclusion, nasal high flow didn't reduce the length of stay, and that was our primary question. And we couldn't identify a subgroup that particularly benefits from the treatment with early high flow. A greater proportion needed ICU admission in the nasal high flow group, but as I showed, this was more driven by the fact having an on-site ICU than not having an on-site. So, that means we need to really think what are we doing within the hospital and how we change our practice using nasal high flow. Nasal high flow doesn't mean the patient is automatically sicker, and then we refer this patient to ICU because our metropolitan and regional hospitals have very clearly shown they keep these patients in their hospital without the transfer. However, tolerance of nasal high flow is still a problem and is probably not that well tolerated and needs to be reconsidered how we're going to use it. But we can certainly conclude with almost no adverse events in our study that nasal high flow can be safely used outside ICUs and as well in hospital without any on-site ICU service. So, thank you very much.

nasal high flow

pediatric hospitals

acute hypoxemic respiratory failure

length of stay

further research