Padmanabhan Ramnarayan. High flow therapy – when and how?, Don't Forget the Bubbles, 2021. Available at:
Isn’t nasal high flow just a fancy name for plain old high flow oxygen? Or is it CPAP-lite? For a therapy that has become so popular in less than a decade, amazingly, there is more opinion sloshing around than proper scientific evidence…
OK, back up, what is Nasal High Flow therapy?
Nasal high flow therapy (NHF), aka high flow nasal cannula therapy (HFNC), is a non-invasive mode of respiratory support, involving the delivery of heated (to 37° C) and humidified (to nearly 100% relative humidity) gas (oxygen and/or medical air) through nasal cannulae at high gas flow rates. What is a “high” gas flow rate is still not uniformly defined (some studies say >2 L/min and others >4 L/min). In physiological terms, to provide the true benefits of “high flow”, the gas flow rate should exceed the patient’s maximal peak inspiratory flow rate (roughly 8-10 x normal minute ventilation).
Makes sense, but what is a child’s peak inspiratory flow rate?
Short answer – it changes according to the age and the extent of respiratory distress. For example, a 4 kg baby breathing at 40 breaths/min and inhaling a tidal volume of 5 ml/kg (=20 ml) would have a minute ventilation of 0.8 L/min and a peak inspiratory flow (PIF) rate of nearly 3 L/min. However, the same infant would have a much higher PIF in respiratory distress. Matching the maximal PIF rate by aiming for roughly 8-10 x normal minute ventilation (in this case, 8-10 x 0.8 L/min = 7-8 L/min) is the key principle of NHF therapy. NB: Magically, the 8 L/min flow rate in this baby is also 2 L/kg/min (more on that later!)
Got it. But when should I start NHF in the ED?
Case 1. A previously healthy 4-month old infant is seen in the emergency department with a two-day history of coryza and poor feeding. On examination, he has mild/moderate subcostal recession and a respiratory rate of 60 bpm. His oxygen saturation in room air is 89%. Would you start nasal high flow?
This baby most likely has mild viral bronchiolitis and the main question is whether to start standard oxygen therapy (SOT) via nasal cannulae or NHF. The most useful clinical outcomes we are interested in are transfer to paediatric intensive care (PICU) and the need for endotracheal intubation.
What does the evidence say?
A recent systematic review (Lin J et al. Arch Dis Child 2019) is an obvious starting point. For the outcome of transfer to PICU, only two RCTs were included (Franklin et al. NEJM, 2018; Kepreotes et al. Lancet 2017). The overall risk ratio was 1.30 (95% CI 0.98, 1.72) indicating no significant difference between NHF and SOT, although there was a tendency to favour SOT.
Similarly, only two RCTs were included for the outcome of intubation (Franklin et al. NEJM, 2018; Yang et al. Chinese Pediatric Emergency Medicine, 2017). The overall risk ratio was 1.98 (95% CI 0.60, 6.56), again with no significant difference between NHF and SOT. So, not much joy from the systematic review…
Considering that Franklin et al dominated the systematic review in terms of sample size, it is useful to look at this RCT in a bit more detail, from a PICO point of view as well as the flow of patients through the RCT.
Population: Infants <12 months of age with bronchiolitis and needing supplemental oxygen
Intervention: NHF at 2 L/kg/min
Control: Standard oxygen therapy
Outcome: Escalation of care due to treatment failure (composite outcome)
A few reflections on the outcomes of infants in this RCT: although nearly double the number of infants randomised to SOT “failed treatment” compared to NHF, it is notable that over 75% of infants randomised to SOT did not “fail”; the majority of those who did “fail” SOT were rescued by NHF; and since NHF “failure” automatically led to PICU transfer, in effect, more infants were transferred to PICU in the NHF group than in the SOT group (12% vs 9%). Essentially, this RCT could be considered a trial of ‘immediate’ NHF versus ‘rescue’ NHF, as covered by us here previously.
Bottom line: There is no advantage to starting NHF as first-line therapy in an infant with mild bronchiolitis. A more clinically and cost-effective strategy would be to use NHF as ‘rescue’ therapy when standard oxygen therapy has failed.
Case 2. An ex-prem born at 24 weeks gestation, now 4 months old, is seen in the emergency department with a 24-hour history of coryza and cough. On examination, he has moderate/severe subcostal recession and a respiratory rate of 80 bpm. His oxygen saturation in room air is 85%. Would you start nasal high flow?
This baby is much sicker, with significant past medical history, and most likely has moderate/severe bronchiolitis. Would NHF be more useful as first-line therapy in this infant, where previously nasal CPAP would have been an option – can NHF be used as ‘CPAP-lite’? A really useful clinical outcome to focus on is endotracheal intubation.
What does the evidence say?
Lin et al summarise the evidence in their recent systematic review. For the outcome of intubation, 4 RCTs were included, but the total number of patients included were low (n=264). Intubation rates were identical in the NHF and CPAP groups (5.3%), with a risk ratio of 0.96 (95% CI 0.35, 2.61). So, there is not much evidence to support the use of NHF compared to CPAP, although quite notably, the rate of adverse events was lower in the NHF group (8% vs 21%).
Bottom line: There is no clinical advantage to starting NHF as first-line therapy in an infant with moderate to severe bronchiolitis to avoid intubation. However, its adverse event profile and tolerability by infants might make NHF more appealing as first-line therapy.
When should I start NHF in the HDU?
Case 3: A 5-year old boy with cerebral palsy and epilepsy is admitted to the paediatric HDU bed with fever, cough and respiratory distress. On examination, he has moderate subcostal and intercostal recession and a respiratory rate of 45 bpm. His oxygen saturation in room air is 88%. Would you start nasal high flow?
In this older child with a complex past medical history, is starting NHF, compared to either standard oxygen or CPAP, beneficial in terms of avoiding the need for endotracheal intubation?
What does the evidence say?
A recent systematic review (Luo J et al. Journal of Pediatrics, 2019) is an obvious starting point. In the comparison of NHF versus SOT, 5 RCTs were included, although 2 were focussed on bronchiolitis (previously covered – Franklin et al and Kepreotes et al). The other 3 RCTs were small (Chisti et al. Lancet, 2015; Ergul et al. Eur J Pediatrics, 2018; Sittikharnka et al. Indian J Crit Care Med 2018) with just 300 patients in total. The overall risk ratio for intubation from these 3 studies alone (calculated specifically for this post) was 0.72 (95% CI 0.38, 1.36). Similarly, in the comparison of NHF versus CPAP, 4 RCTs were included but 2 were in bronchiolitic infants (covered earlier). The other two RCTs (Ramnarayan et al. Crit Care 2018; Chisti et al. Lancet 2015) included just 187 children. The overall risk ratio for intubation based on these two RCTs (calculated for this post) was 2.14 (95% CI 0.93, 4.92) indicating a tendency for a higher intubation rate with NHF in older children.
Bottom line: In an older child, intubation was not less frequent when NHF was used compared to SOT as first line therapy. There was a tendency for NHF to be associated with a higher intubation rate compared to CPAP.
Great – what is the best way to provide NHF?
Starting gas flow rate
Milesi et al showed in physiological studies in infants aged <6 months with bronchiolitis that the work of breathing is reduced considerably when the gas flow rate is set at nearly 2 L/kg/min. In their cohort of 21 infants (mean weight 4.3 kg), the measured work of breathing was lowest at a flow rate of 7 L/min (compared to 1, 4 and 6 L/min). Similarly, in children up to the age of 3 years with pneumonia, work of breathing was reduced by nearly 20% at a flow rate of 1.5 L/kg/min compared to 0.5 L/kg/min (Weiler et al. Journal of Pediatrics 2017). Usual adult flow rates range from 50-60 L/min.
In summary, the optimal gas flow rate does not increase in a linear fashion with increasing age/weight, instead it goes from nearly 2 L/kg/min in infancy to nearly 1 L/kg/min in young adults.
RCTs of different starting flow rates
There have been two RCTs comparing NHF flow rates in bronchiolitis (Yurtseven A et al. Ped Pulm 2019; Milesi et al. Intens Care Med 2018). In the former, 1 L/kg/min (n=88) was compared to 2 L/kg/min (n=80) in infants <24 months with clinical severe bronchiolitis presenting to the emergency department. The primary outcome was ‘treatment failure’ within 24 hours. There was no significant difference in treatment failure between the two flow rates (1 L/kg/min: 11.4%; 2 L/kg/min: 10%). The second RCT compared 2 L/kg/min (n=142) with 3 L/kg/min (n=144) in infants aged <6 months with moderate/severe bronchiolitis. The primary outcome was treatment failure within 48 hours. There was no significant difference in treatment failure between the two groups (2 L/kg/min: 38.7%; 3 L/kg/min: 38.9%).
A useful chart with suggested starting flow rates based on weight is used in the ongoing FIRST ABC clinical trial of NHF versus CPAP.
Nasal cannula size
There are different nasal cannula sizes available based on the manufacturer. The general rule of thumb is that the cannula prongs should be no more than 50% of the diameter of the nostril to avoid inadvertent occlusion of the nasal passages. It is also advisable to start the gas flow rate at a low rate and then increase gradually over 10-15 min to avoid patient discomfort. Pacifiers may be useful in babies to prevent mouth opening.
There are no RCTs comparing weaning strategies for NHF. Clinical practice is also highly variable – in a global survey of practice in over 1000 PICU professionals by Kawaguchi et al, 68% weaned the FiO2 first to a threshold value (e.g. 0.40) and then weaned the flow rate gradually, 11% weaned the FiO2 first to a threshold value (e.g. 0.40) and then stopped NHF, and 4% weaned the flow rate alone without weaning the FiO2. The FIRST ABC RCT algorithm for the weaning of NHF provides a weight-based approach to a one-step weaning process and suggested clinical thresholds for weaning and stopping NHF.
The take homes
- Nasal high flow is a form of non-invasive respiratory support that sits somewhere between standard oxygen therapy and nasal CPAP.
- In infants with mild bronchiolitis, there is no clinical (or cost) benefit in starting NHF as first-line treatment – rather, NHF is best used as a ‘rescue’ therapy after standard oxygen.
- In infants with moderate/severe bronchiolitis, NHF may be a useful first-line therapy owing to its ease of use and since it is better tolerated by infants, however there is no clinical benefit compared to nasal CPAP.
- In older children with respiratory failure, there is little RCT evidence to guide practice – however, there is no clear benefit of starting NHF over and above standard oxygen. NHF may be associated with a higher intubation rate compared to CPAP in older children.
- There is no RCT evidence to support either 1, 2 or 3 L/kg/min NHF flow rates in infants with bronchiolitis; however, physiological evidence suggests that nearly 2 L/kg/min is associated with reduction in work of breathing. Suitable flow rates in older children approximate 1.5 L/kg/min and in young adults, 1 L/kg/min.
- There is no RCT evidence to support one way of weaning over another – the most common practice seems to be to reduce FiO2 to below 0.40, followed by a reduction in the NHF flow rate.
- Ongoing RCTs such as the FIRST ABC trial will help address the question whether NHF is non-inferior to CPAP in critically ill children.