52 in 52 – #28: A Randomized Trial of High-Flow Oxygen Therapy in Infants with Bronchiolitis

Clinical question:

In infants (less than 12 months old) with bronchiolitis and hypoxemia (saturation less than 92-94%) in the emergency department or pediatric ward, does early initiation of high-flow nasal cannula versus standard nasal cannula oxygen therapy lead to fewer instances of escalation of care?

Study design:

• Block sequence assignment RCT

PICO:

Population:

• 17 hospitals (6 tertiary and 11 regional) in Australia and New Zealand, where the standard of care was early high-flow nasal cannula (HFNC).
• Inclusion Criteria:
  • Infants < 12 months old
  • Clinical signs of bronchiolitis
  • Hypoxemia (defined differently by center; SpO2 < 92% at tertiary centers versus SpO2 < 94% at regional hospitals)
  • Patient was located in the emergency department or general pediatric ward
• Exclusion Criteria:
  • Critically ill children with an immediate need for respiratory support and/or ICU admission,
  • Infants with a history of cyanotic heart disease
  • Infants with craniofacial malformation
  • Infants with a basilar skull fracture
  • Infants who receive oxygen therapy at baseline, and/or
• 2,217 infants were eligible for inclusion; 1,638 underwent randomization, of which 210 declined consent.
• Powered to 90% at a type I error rate of 0.05 with assumptions of failure (requiring escalation) rate of 10% in the standard nasal cannula group and 5% with an assumption of 10-20% loss to follow-up.

Intervention: 

• Initiation of early high-flow nasal cannula at 2 liters/kilogram of body weight/minute.
• Titration of the fraction of inspired oxygen (FiO2) 21%-40% to maintain SpO2 of 92%-98% (94-98% at regional centers).
• High flow discontinued after > 4 hours at 21% FiO2.
• Care was escalated if 3 of the following 4 were present:
  • Heart rate was the same or increased after initiation of therapy
  • Respiratory rate was the same or increased after initiation of therapy
  • Oxygen requirement exceeded FiO2 of 40%
  • Hospital early warning tool triggered medical review for escalation of care
• Clinicians could decide to preference to escalate regardless of the above objective metrics.

Control:

• Initiation of standard nasal cannula therapy.
• Titration from 0 to 2 liters per minute to maintain SpO2 of 92%-98% (94-98% at regional centers).
• Care was escalated if 3 of the following 4 were present:
  • Heart rate was the same or increased after initiation of therapy,
  • Respiratory rate was the same or increased after initiation of therapy,
  • Oxygen requirement 2 liters per minute,
  • Hospital early warning tool triggered medical review for escalation of care.
• Clinicians could decide to preference to escalate regardless of the above objective metrics

Outcome:

• Authors report that HFNC group required significantly less escalation of therapy (87 of 725) than the standard nasal cannula group (167 of 726)
• Every control group member who required escalation was escalated to HFNC; 102 were stabilized with HFNC while only 65 required further escalation.
• 53 patients in the intervention group and 115 patients in the control group were escalated by their clinicians despite the objective metrics for escalation not being met.
• No significant differences in secondary outcomes, including:
  • Duration of hospital stays
  • The proportion of transfers or upgrades to ICUs
  • Duration of ICU stays
  • Duration of oxygen therapy
  • Intubation rates
  • Serious adverse events (defined as fatal, life-threatening, permanently disabling, or resulting in a prolonged hospital stay)

• Though the control group required more “escalation of care,” this escalation was uniformly to the level of care the other group was receiving at baseline; no infants receiving standard nasal cannula required direct escalation to BPAP, CPAP, or mechanical ventilation.
• In fact, there were numerically fewer infants in the control group who required escalation beyond HFNC than the number of infants escalated from the intervention group (i.e., HFNC used as rescue therapy seemed just as efficacious (if not more) in preventing further escalation as early HFNC).
• Thus, while the primary outcome was “escalation of care,” the definition of “escalation of care” varied between the groups, rendering it an unhelpful primary outcome.
• Meanwhile, there were no important differences in secondary outcomes, which overall were more directly comparable and patient-centered.
• That so many patients in the control group were escalated to HFNC despite objective metrics not requiring escalation could be evidence that clinicians were biased by the erstwhile institutional standard of care and/or that the objective metrics used by the study should have been refined (for example, to include presence/improvement in retractions, nasal flaring, etc.)
• This study has several weaknesses, including the likelihood for selection bias (many eligible infants were never screened/randomized), detection bias (clinicians could escalate care regardless of objective metrics), that there were differences by the institution in the definition of hypoxemia, and that the interventions were necessarily unblinded.
• The authors underestimated the failure rate in their power calculations.

My take:

• HFNC can be reasonably used as a rescue therapy rather than a standard therapy in hospitalized infants diagnosed with bronchiolitis and hypoxemia.
• Although it is not wrong to use HFNC from the outset in infants suffering from bronchiolitis, it may result in a misuse of resources.
• If the patient fails other therapies or presents with respiratory distress, consider HFNC.

For more on bronchiolitis, see this Pediatric Small Talks post.

References:

1. Franklin D, Babl F, Schlapbach L, et al. A Randomized Trial of High-Flow Oxygen Therapy in Infants with Bronchiolitis. N Engl J Med. 2018;378(12):1121-1131. Doi: 10.1056/NEJMoa1714855