Many, many children will suffer from the symptoms of bronchiolitis. Some estimate that it will effect >90% of infants and toddlers under the age of 2 years. We are often at a loss for what to do for some of these babies. They are tachypneic, with retractions and signs of respiratory distress. We want to help them, and when something new and promising comes along we latch onto it. Suffice to say, high flow, has been latched onto in a major way. I have seen it used appropriately, inappropriately, just because care team member were worried that the baby “might get worse,” because that’s what the floor will do anyway so why don’t we just start it, and more. Per some studies HFNC is estimated to be 16 times more expensive than standard nasal cannula.
Therefore, the purpose of this post is to look at the literature on the use of high flow nasal cannula in children with bronchiolitis. This is especially pertinent here in the Spring of 2020 during the COVID-19 pandemic, as there is ongoing discussion about whether or not HFNC is aerosolizing.
How is HFNC being used?
Broadly and without consensus… But in all seriousness, HFNC is used in children with bronchiolitis who have moderate to severe increased work of breathing, refractory hypoxia, and/or episodes of apnea. In general initial flow rates are based on age and patient size, and are typically 1 to 2 L/kg.
What are the proposed physiologic benefits?
Milési et al noted that 2L/kg/minute resulted in mean pharyngeal pressures of >4 cm H2O and improved breathing. in adults (who can keep their mouth closed) there is a linear relationship between flow rate and pharyngeal pressure, from 1-4 cm H2O.
Reduced upper airway resistance
The flow rates of HFNC are greater than that of normal inspiratory pressure. Studies of pulmonary function tests and other dynamic measurements of diaphragm action have been used to assess this.
Washout of dead space in the nasopharynx
It is thought that HFNC washes out expired gas in the nasopharynx, and creates a reservoir of richly oxygenated air.
The warm, humidified air helps too
Dysart et al noted that the warm, humidified air of HFNC reduces oxygen demand in children who have tachypnea and who are rapidly turning over their tidal volume. It may also lead to decreased airway reactiviy and bronchospasm.
Are there any potential harms?
The main risks of HFNC include; gastric distention and resultant poor feeding, barotrauma/pneumothorax, an the increased cost.
What’s the evidence?
Let’s take a look at some of the many recent studies on HFNC in bronchiolitis. I will briefly comment on the significance and impact to the broader conversation to the best of my abilities. Note that there are substantially more observational studies than randomized controlled trials. I also can’t claim to have included every article – I just tried to pick the most relevant ones that will give you a sense of where we have been, and where we are currently. Many studies also compare nasal CPAP to HFNC. I have decided to focus on standard nasal cannula (NC) versus HFNC – since this is the question that arises the most in current EDs.
This is where the bulk of the evidence lies unfortunately.
This retrospective review of 46 patients noted that respiratory scores improved. they also observed better looking X-Rays and that 5/46 needed mechanical ventilation.
This retrospective study compared groups of patients before and after the availability of HFNC between 2006-09 at a single site. The authors noted that “after controlling for age, month of admission, type of respiratory illness, and severity of illness, there was an 83% reduction in the odds of intubation in the PED in patients with HFNC availabel versus not (odds ratio, 0.17; 95% CIl, 0.06-0.50; P = 0.001).” They noted no significant difference in mortality or PICU length of stay after the introduction of HFNC.
Anecdotally the above paper was the first study that lit the HFNC flame in the Pediatric ED.
This retrospective cohort in two hospitals included 498 children under two with bronchiolitis and other respiratory diseases that got HFNC and were seen between June 2011 and September 2012. They concluded that a “diagnosis of bronchiolitis was observed to be protective with respect to intubation (OR, 0.40; 95% CI, 0.17-0.96).” However, this study did not compare HFNC with any other intervention.
This prospective, multicenter observational study included patient data from 16 children’s hospitals between 2001 and 2010. The main take home was that there was a large amount of institutional variability in the use of HFNC and that any conclusions about best practices could not be made.
This was a retrospective, pre-post intervention study of 1,937 children with bronchiolitis seen at a single children’s hospital between 2010 and 2014. They split the study into 2010-12 (pre HFNC) and 2012-14 (after HFNC) and noted that though HFNC use increased, total hospital length of stay (P = .48), PICU length of stay (P = .06), or rate of PICU transfer (P = .97) were not signifincatly different. the authors noted that “there was also no difference in intubation rate or 30-day readmission between the 2 groups.”
Randomized Controlled Trials
This study of 201 Australian children <24 months with moderate bronchiolitis in a single center. The main outcome was time receiving oxygen, and there was no difference in the HFNC and standard cannula groups (20 hours in the HFNC group vs 24 hours in the standard cannula group; hazard ratio, 0.9; 95% CI, 0.7-1.2). The standard nasal cannula did have a higher rate of treatment failure (33%, n=33 for NC versus 14%, n=14 for HFNC0 – as defined by moving to HFNC and/or going to the PICU. Interestingly, both NC and HFNC had the same proportion going to the PICU.
This larger study was also conducted in the southern hemisphere, and included 1,472 children from Australia an New Zealand. Higher flow rates than Keproetes (2 L/kg) were used in infants, Overall the authors noted a higher rate of treatment failure in the standard nasal cannula group (NC – 23% [167 of 733] versus HFNC – 12% [87 of 739]). Again, NC failure meant that the child transitioned to HFNC. 9% of the NC kids went to the PICU versus 12 % of the HFNC. There were no differences in overall hospital length of stay or duration of oxygen therapy.
Both of these RCTs allowed for crossover. This is incredibly important because in reality, they were comparing using HFNC early, versus using it if NC failed. They both appear to be equal in terms of their outcomes. HFNC also costs a whole lo more to administer…
Meta Analyses and Systematic Reviews
They planned to include RCTs and didn’t find any. So this initial meta analysis said that there was (at the time of publication) no evidence about the effectiveness of HFNC.
This review included 9 RCTs and 2,121 children. Overall they noted that HFNC was safe, and that there was a decreased risk of treatment failure in the HFNC group versus standard NC or nasal CPAP – (RR 0.50, 95% CI 0.40 to 0.62, p<0.01). There was also no There was “no significant difference in length of stay in hospital (LOS), length of oxygen supplementation (LOO), transfer to intensive care unit, incidence of intubation, respiratory rate, SpO2 and adverse events” when HFNC was compared with NC and nasal CPAP.
I think that the next meta analysis better illustrates why some RCTs help define these outcomes better than others.
This meta analysis included 8 randomized controlled trials with 2,259 patients and compared HFNC (or nasal CPAP) with standard nasal cannula in bronchiolitis (5 studies) and 3 trials of severe pneumonia (as defined by WHO criteria). The authors concluded that the incidence of treatment failure was reduced overall (see the forest plot below). Note that treatment failure was defined a bit differently by each study; but generally the persistence of hypoxia, inability to reduce respiratory rate or heart rate by 20%, increased apnea episodes, improved respiratory scores, and escalation in respiratory support (including going to HFNC and mechanical ventilation).
One could argue (persuasively) that Kepreotes and Franklin are the two most relevant trials here. As noted earlier, nasal cannula treatment failure lead to HFNC. So really, these two compared early versus rescue use of HFNC. Both also noted no difference in duration of oxygen need, length of stay, ICU transfer, or other adverse events. So, starting HFNC early prevents you from starting it later – but the outcome didn’t change otherwise. Weird huh?
OK, so clearly HFNC isn’t useless – but most of the evidence suggests that starting it early prevents you from starting it later on in the hospital course, but not necessarily intubation risk and hospital length of stay. So, don’t use HFNC if you think that it will broadly reduce the number of intubations or PICU transfers you expect to see at your hospital. i.e. don’t make it your standard treatment for all bronchiolitis patients.
Do start high flow nasal cannula if…
- An infant has significant respiratory distress – fatiguing, head bobbing, grunting, nasal flaring, or frequent apneic episodes – as it can be helpful.
- If you and your team feel that in this particular patient it might buy you some time and you feel that they need positive pressure support
Don’t start high flow…
- If the patient is already apneic, gray or blue – just intubate them
- Just because the floor asks for it
- Because you think you might need it in the future. In those cases, if you think that child needs support try 100% FiO2 via a standard nasal cannula. Seriously, it will help. It helped before HFNC was available
Ultimately we need better RCTs, preferably multicenter ones, that will help us develop more consistent protocols and really understand if high flow works like we think, and sometimes hope it does. Especially now, with the COVID-19 pandemic flexing its muscles we also need to be careful about potential aerosolizing procedures.
Milesi et al. Is treatment with a high flow nasal cannula effective in acute viral bronchiolitis? A physiologic study. Intensive Care Med. 2013 Jun;39(6):1088-94. doi: 10.1007/s00134-013-2879-y. Epub 2013 Mar 14.