Date: April 20, 2023

Reference:  Franklin D, et al. Effect of early high-flow nasal oxygen vs standard oxygen therapy on length of hospital stay in hospitalized children with acute hypoxemic respiratory failure: the PARIS-2 randomized clinical trial. JAMA 2023

Guest Skeptic: Dr. Michael Falk is a Pediatric Emergency Medicine attending at Mount Sinai Medical Center and Associate Professor of Emergency Medicine at the Icahn School of Medicine in New York. He is interested in simulation and medical education.

Dr. Michael Falk

Case: A 14-month-old boy presents to the emergency department (ED) with two days of upper respiratory symptoms and respiratory distress.  He has a cough, fever of 38.5°C, runny nose and increased work of breathing that started today.  He is breathing at rate of 48 breaths per minutes with intercostal retractions and oxygen saturation of 88%.  His lung exam reveals bilateral wheezing, rhonchi but no focal findings. He is drinking well, and parents report normal urine output.  He is suctioned and given trial of Beta agonist because has a history of eczema and a sibling with asthma with no change.  Despite your interventions, he continues to breathe rapidly with an oxygen saturation of 89% on room air.  

You are working with an eager medical student, and she asks, “Should we start high flow nasal cannula (HFNC) at 2L/kg/min and admit the patient?”

Background: Respiratory illnesses remain one of the most frequent causes of admission for children less than 5 years.  Some of these illnesses result in acute hypoxemic respiratory failure. Historically, there were not many treatments for these children, and they were admitted for observation or intubated and started on mechanical ventilation. High flow nasal cannula (HFNC) started gaining popularity in the early 2000s as an option of noninvasive ventilation.  

We have covered the use of HFNC in pediatrics a few times on the SGEM including SGEM #228 and SGEM #379. Previous research has shown that HFNC can lower the rate of escalation of care but showed no impact on admission to the intensive care unit (ICU) or length of stay [1]. 

Clinical Question:  Does the early use of HFNC reduce the length of hospital stay in pediatric patients with acute hypoxemic respiratory failure compared with standard oxygen therapy?

Reference:  Franklin D, et al. Effect of early high-flow nasal oxygen vs standard oxygen therapy on length of hospital stay in hospitalized children with acute hypoxemic respiratory failure: the PARIS-2 randomized clinical trial. JAMA 2023

  • Population: Children aged 1 to 4 years of age who presented across 14 emergency departments in Australia and New Zealand requiring hospital admission for acute hypoxemic respiratory failure
    • Exclusion: There was a long list of exclusion criteria that you can find in the supplemental material, but these included craniofacial abnormalities, upper airway obstruction, cyanotic heart disease, tracheostomies, apneas, immediate high-level care in the ICU or required noninvasive or invasive mechanical ventilation.
  • Intervention: HFNC at differing rates depending on weight

  • Comparison: Oxygen via nasal cannula at 2L/min or by face mask up to 8L/min 
  • Outcome: 
    • Primary Outcome: Length of hospital stay defined as time from randomization to time of hospital discharge/death.
    • Secondary Outcomes:
      • Length of oxygen therapy from the time of randomization
      • Length of hospital stay starting from arrival in the ED
      • Proportion of children requiring a change in therapy on the general ward
      • Proportion of children that required ICU admission or transfer to hospital with a pediatric ICU
      • Proportion of children that required escalation of care to noninvasive or invasive ventilation
      • Adverse events
      • Tolerance of intervention
      • Clinical triggers that warranted a change in that child’s care
  • Type of Study: Multicenter, randomized clinical trial 

Authors’ Conclusions: Nasal high-flow oxygen used as the initial primary therapy in children aged 1 to 4 years with acute hypoxemic respiratory failure did not significantly reduce the length of hospital stay compared to standard oxygen therapy.”

 Quality Checklist for Randomized Clinical Trials:

  1. The study population included or focused on those in the emergency department.  Yes 
  2. The patients were adequately randomized. Yes
  3. The randomization process was concealed. No
  4. The patients were analyzed in the groups to which they were randomized. Yes
  5. The study patients were recruited consecutively (i.e. no selection bias). No
  6. The patients in both groups were similar with respect to prognostic factors. Yes
  7. All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No
  8. All groups were treated equally except for the intervention. Unsure
  9. Follow-up was complete (i.e. at least 80% for both groups). Yes
  10. All patient-important outcomes were considered. Yes
  11. The treatment effect was large enough and precise enough to be clinically significant. Yes
  12. Financial conflicts of interest. Many of the authors had received grants from various governmental organizations or biomedical companies including the company that supplied the OptiFlow equipment for the study. They declare that the funder/sponsor had no role in designing or conducting the study, the data, or the manuscript. 

Results: 1,567 children were randomized. 782 were included in the HFNC group and 785 were included in the standard oxygen therapy group. They did lose a few children in each group because parents either declined deferred consent, they were unable to obtain consent, or the parents withdrew consent. Median age in both groups was 1.9 years and overall characteristics were very similar.

Key Results: Early initiation of HFNC did not significantly reduce length of stay for pediatric patients presenting with acute hypoxemic respiratory failure. 

Primary Outcome: Length of hospital stay was significantly longer for patients who received HFNC compared to those treated with standard oxygen therapy.  

High-flow oxygen: 1.77 days (IQR 1.03-2.80)

Standard oxygen: 1.50 days (IQR 0.85-2.44)

Adjusted Hazard Ratio (95% CI): 0.83 (0.75 to 0.92)

This was not impacted by presence/absence of wheezing, or obstructive disease

Secondary Outcomes:

Median length of stay from presentation to the ED and median length of oxygen therapy was also longer in the HFNC group.

A higher proportion of patients in the HFNC group required escalation to the ICU.

42.9% of children in the high-flow oxygen group were switched to standard oxygen while 18.5% of children in the standardized oxygen group were switched to high-flow oxygen.

1) Selection Bias: There were 1,348 parents who were not approached for enrollment in this study. One of the criteria for inclusion in the study was the need for admission to the hospital. This is a subjective decision and may be physician dependent.

One group of the patient who were excluded from this study were those who required immediate higher-level care in the ICU. Depending on where you practice, there are some institutions where the initiation of HFNC automatically lands the patient in the ICU. Individual clinicians may have differing thresholds for initiation of HFNC for pediatric patients in respiratory distress. It would seem like the patients in this study were kind of sick but not super sick. 

2) Oxygen Saturation Thresholds:One of the criteria for which patients were enrolled was based on their oxygen saturations. They had two different thresholds depending on the study site of 90% or 92%. They did watch these children for 10 minutes to see if it was sustained but any child with an oxygen saturation of less than 85% was started immediately on supplemental oxygen. After implementing treatment, they also had target oxygen saturations of 90/92% to 98%. This is a monitor-oriented outcome, and we can already see that there are different standards depending on the institution for what the lowest threshold of oxygen saturation should be before starting oxygen supplementation.

Children with bronchiolitis can have transient desaturations into the 70s or 80s without significant clinical consequence.[2] Additionally, research suggests that there may be racial bias in the use of pulse oximeters as the readings in darker-skinned individuals may be inaccurate. [3-5] In summary, using oxygen saturations alone to make clinical decisions is imperfect. [6]

3) Weaning Respiratory Support: They made an interesting decision here when it came to weaning respiratory support. While they were able to adjust the fraction of inhaled oxygen (FiO2) based on oxygen saturations, the flow itself was never changed. When a patient was taken off of high-flow oxygen therapy, it is possible that they were weaned immediately from 40L/min. 

While weaning practices vary, this feels a bit aggressive. It is uncertain how this practice could have impacted their results. Could this method have caused more patients to stop and then restart HFNC, prolonging the hospital length of stay and time on oxygen compared to a more gradual weaning of the flow? Additionally, because the treating clinicians could not be blinded to the intervention, could there have been some bias that children on HFNC were maybe sicker than those on standard therapy, leading to hesitation to wean?

4) Clinical Judgement and Confounding: By dividing both the control and the treatment in groups into those who had “wheezing” or not, they also controlled for a potential confounding variable in their design.  One could reasonably hypothesize that patients with bronchiolitis, pneumonia might behave differently than those who had a reactive airway disease picture.

This makes their results more applicable to clinical practice given the heterogeneity of the underlying conditions that were included. On the other hand, because they included so many different conditions, and patients were treated the “hospital’s standard ED management,” that may introduce some confounding based on what treatments the child received prior to enrollment.

5) Crossover Between Groups: 42.9% of children in the HFNC group crossed over and received standard oxygen therapy while 18.5% of the children in the standard oxygen therapy group crossed over into the HFNC group. This decision could have been up to the clinicians (which we think is appropriate) or changes in vital signs, clinical exam, or patient tolerance. 

It is worth noting that more patients in the HFNC were switched due to intolerance to the therapy compared to standard oxygen therapy. This is an important patient-oriented outcome to consider when deciding which therapy to initiate. Maybe this switch from HFNC to standard oxygen therapy also played a role in the increased length of stay for the HFNC group.

Comment on Authors’ Conclusion Compared to SGEM Conclusion: We agree with the authors’ conclusions. 

SGEM Bottom Line: HFNC significantly increased the length of stay for pediatric patients with acute hypoxemic respiratory failure compared to standard oxygen therapy.

Case Resolution: You tell the medical student that you agree with her conclusion that the child requires admission. However, given the results of the recent PARIS-2 trial you would not start high-flow oxygen therapy at this time. The patient is admitted to the general pediatric floor with oxygen via face mask.

Clinical Application: High-flow oxygen therapy is very popular, but this isn’t the first time that medicine has embraced an intervention with limited evidence. It does not seem to decrease length of stay or escalation to the ICU or decrease risk of requiring non-invasive/invasive ventilation. 

We are overusing HFNC and really need to be thoughtful about what we are treating. Is it severe respiratory distress or our own discomfort? The indiscriminate use of HFNC can be worse for patients as it extends their length of stay. For those practicing in the emergency department, we are often the ones making that initial decision. [7] So, for pediatric patients with mild to moderate hypoxemic respiratory failure, start them on standard oxygen therapy rather than HFNC.

What Do I Tell the Patient/Family?  Your child has something called bronchiolitis and is breathing very fast. His oxygen level on the monitor is also a bit low which can happen in this disease. Many of these children improve on their own with time, but I believe he need some oxygen to help him out right now. We can start with oxygen delivered by face mask and have you stay in the hospital for close monitoring. 

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics Guide to Emergency Medicine.

Other FOAMed:

  1. First10EM: Blow some O2 up that nose?
  2. Don’t Forget the Bubbles: Paris in Autumn
  3. Don’t Forget the Bubbles: High Flowing Controversy
  4. Rebel EM: The PARIS Trial 


  1. Franklin D, Babl FE, Schlapbach LJ, et al. A randomized trial of high-flow oxygen therapy in infants with bronchiolitis. N Engl J Med. 2018;378(12):1121-1131.
  2. Principi T, Coates AL, Parkin PC, Stephens D, DaSilva Z, Schuh S. Effect of oxygen desaturations on subsequent medical visits in infants discharged from the emergency department with bronchiolitis. JAMA Pediatr. 2016;170(6):602-608.
  3. Sjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial bias in pulse oximetry measurement. N Engl J Med. 2020;383(25):2477-2478.
  4. Valbuena VSM, Seelye S, Sjoding MW, et al. Racial bias and reproducibility in pulse oximetry among medical and surgical inpatients in general care in the Veterans Health Administration 2013-19: multicenter, retrospective cohort study. BMJ. 2022;378:e069775.
  5. Ruppel H, Makeneni S, Faerber JA, et al. Evaluating the accuracy of pulse oximetry in children according to race. JAMA Pediatr. Published online March 20, 2023:e230071.
  6. Schuh S, Freedman S, Coates A, et al. Effect of oximetry on hospitalization in bronchiolitis: a randomized clinical trial. JAMA. 2014;312(7):712-718.
  7. Byrd C, Noelck M, Kerns E, et al. Multicenter study of high-flow nasal cannula initiation and duration of use in bronchiolitis. Hosp Pediatr. Published online March 20, 2023:e2022006965.