High-flow nasal oxygen

Contents

Overview

High-flow nasal oxygen (HFNO) delivers heated (37°C) and humidified oxygen at flows of up to 60 L/min via wide-bore nasal cannulae. It provides a precise, adjustable FiO2 (0.21–1.0) and generates a modest positive end-expiratory pressure effect, while improving patient comfort compared with standard high-flow face masks.

HFNO is now established as a first-line respiratory support modality for acute hypoxaemic respiratory failure in adults and is increasingly used in the post-extubation period.

Mechanisms of Benefit

Accurate FiO2 delivery: At flows of 40–60 L/min, delivered flow exceeds the patient's peak inspiratory flow rate, preventing entrainment of room air and providing a consistent inspired oxygen fraction.

Nasopharyngeal dead space washout: Continuous high-flow gas clears expired carbon dioxide from the nasopharynx and anatomical dead space, reducing re-breathing and effectively reducing physiological dead space. This contributes to improved alveolar ventilation for a given respiratory effort.

Modest PEEP effect: High flow generates approximately 0.35–0.5 cmH2O of positive pressure per 10 L/min of flow in adults breathing through the mouth closed. At 60 L/min, this equates to 2–3 cmH2O PEEP — clinically modest but potentially significant in maintaining airway opening and alveolar recruitment.

Heated and humidified gas: Reduces upper airway resistance and improves mucociliary clearance. Avoids the desiccation of secretions and bronchospasm associated with cold, dry high-flow oxygen.

Patient comfort: Better tolerated than non-rebreather masks, particularly over prolonged periods. Patients can speak and eat while receiving HFNO.

Practical Use

Starting settings:

  • Flow: start at 30–40 L/min; increase in increments to a maximum of 60 L/min based on respiratory rate, work of breathing, and SpO2
  • FiO2: titrate to SpO2 target (94–98% in most patients; 88–92% in patients at risk of hypercapnic respiratory failure)
  • Temperature: 37°C (34°C if the patient finds 37°C intolerable)

Monitoring: Reassess frequently in the first 1–2 hours. Track respiratory rate, work of breathing, SpO2, and the trajectory of FiO2 requirements. An ABG at 1–2 hours guides decision-making about escalation.

Escalation triggers: Increasing FiO2 requirement, persistent or worsening tachypnoea (RR >30), increasing use of accessory muscles, hypercapnia on ABG, clinical signs of fatigue, reduced consciousness, or haemodynamic instability all warrant consideration of escalation to NIV or invasive ventilation.

ROX Index

The ROX (Respiratory rate-OXygenation) index predicts the success of HFNO and guides decision-making about escalation to intubation.

ROX index = (SpO2 / FiO2) / Respiratory rate

Key thresholds (from Roca et al., Intensive Care Medicine 2016 and subsequent validation):

  • ROX >4.88 at 12 hours: associated with success (avoidance of intubation)
  • ROX <3.85 at any assessment: high risk of HFNO failure, consider early intubation
  • Serial measurements are more informative than a single time point

The ROX index is a practical tool for structured reassessment and explicit decision-making. It is not a substitute for clinical judgement — a patient may have a ROX above the threshold but be visibly deteriorating.

Key Trials

FLORALI (Frat et al., NEJM 2015): Randomised 310 patients with acute hypoxaemic respiratory failure (non-hypercapnic, PaO2/FiO2 <300 mmHg) to HFNO, standard oxygen via non-rebreather mask, or non-invasive ventilation. Intubation rates were non-significantly lower with HFNO. At 90 days, HFNO was associated with lower mortality than NIV or standard oxygen (12% vs 22.8% vs 12%, favouring HFNO; p=0.02 for the overall comparison). The trial established HFNO as the preferred first-line modality in acute hypoxaemic respiratory failure without hypercapnia.

Post-extubation HFNO (Hernandez et al., JAMA 2016): Randomised 527 low-risk extubated patients to HFNO vs conventional oxygen. HFNO reduced post-extubation respiratory failure (8.3% vs 14.4%) and re-intubation (4.9% vs 8.3%). A subsequent trial in high-risk patients showed HFNO was non-inferior to NIV.

HFNO in immunocompromised patients: Evidence is weaker. Early retrospective data suggested benefit, but the FLOWOX trial did not show a mortality advantage for HFNO over standard oxygen in immunocompromised patients.

COVID-19

HFNO was widely used during the COVID-19 pandemic for patients with type 1 respiratory failure due to SARS-CoV-2 pneumonia. It reduced intubation rates compared with standard oxygen when used in appropriately monitored settings. Concerns about aerosol generation in COVID-19 led to recommendations to place a surgical mask over the HFNO cannulae and cohort patients appropriately, though evidence that HFNO significantly increases room contamination compared with unmasked breathing is limited.

Limitations and Contraindications

Hypercapnic respiratory failure: HFNO does not reliably support ventilation. In patients with acute hypercapnic exacerbation of COPD or other conditions with CO2 retention, NIV is preferred. HFNO may be used as a ceiling of care in patients unsuitable for NIV, but should not delay NIV in patients who would benefit.

Risk of delayed intubation: Patients receiving HFNO may maintain acceptable SpO2 while working hard, masking impending respiratory failure. Close monitoring with serial ABGs and ROX index calculations is essential. The maxim "don't let HFNO lull you into false security" is particularly relevant in immunocompromised patients with PCP, in whom abrupt deterioration can be catastrophic.

Claustrophobia and patient compliance: A small proportion of patients cannot tolerate the flow or the device despite acclimatisation.

Absent CPAP equivalent: The PEEP generated is modest and insufficient to manage significant pulmonary oedema alone.

Viva Questions

What are the mechanisms by which high-flow nasal oxygen improves oxygenation?

HFNO improves oxygenation through several complementary mechanisms. First, it provides a reliable high FiO2 because the delivered flow (up to 60 L/min) exceeds the patient's peak inspiratory demand, preventing room air entrainment — something standard masks cannot achieve at lower flow rates. Second, the continuous high-flow gas washes out nasopharyngeal dead space, reducing re-breathing of CO2 and improving the efficiency of each breath. Third, the high flow generates a modest positive airway pressure of approximately 2–3 cmH2O at 60 L/min, which helps maintain alveolar patency and recruits atelectatic areas. Fourth, the heated and humidified gas reduces upper airway resistance and supports mucociliary clearance. The combination of these effects reduces work of breathing, reduces respiratory rate, and improves gas exchange, and is reflected in studies showing lower PaCO2 and improved oxygenation indices with HFNO compared with standard oxygen.

What is the ROX index and how is it used clinically?

The ROX index is calculated as (SpO2/FiO2) divided by respiratory rate. It quantifies the relationship between the oxygenation achieved and the ventilatory work required to achieve it. A high ROX indicates that the patient is achieving good oxygenation with little respiratory effort — a marker of HFNO success. A low ROX indicates poor oxygenation relative to respiratory effort, predicting HFNO failure and the need for escalation. The validated thresholds are a ROX above 4.88 at 12 hours predicting success, and a ROX below 3.85 at any time point predicting failure with a high specificity. In clinical practice, the ROX is used as part of a structured reassessment protocol: calculated at 1–2 hours and 12 hours after starting HFNO, alongside clinical assessment of respiratory rate, accessory muscle use, and ABG results. A falling ROX or a value below the failure threshold should prompt preparation for escalation to NIV or intubation, and should not be ignored even if SpO2 is acceptable.

What does the FLORALI trial tell us about the role of HFNO in acute hypoxaemic respiratory failure?

The FLORALI trial randomised 310 patients with acute non-hypercapnic hypoxaemic respiratory failure to HFNO, standard non-rebreather mask oxygen, or intermittent NIV. The primary endpoint of intubation at 28 days did not differ significantly between groups. However, the 90-day mortality was significantly lower in the HFNO group compared with NIV (12% vs 22.8%), and lower than standard oxygen, though the comparison with standard oxygen was not statistically significant in isolation. The trial was important for several reasons: it was the first adequately powered trial to compare these modalities head to head, it established that HFNO is at least as safe as NIV in this population, and the 90-day mortality signal — even if unexpected — shifted clinical practice significantly towards HFNO as the default first-line device for acute type 1 respiratory failure. Limitations include that it was not blinded, the NIV regimen was intermittent rather than continuous, and the population was selected (required PaO2/FiO2 <300 and no hypercapnia). The trial does not address hypercapnic respiratory failure, in which NIV remains the evidence-based intervention.