By Dr. Sabrina Berdouk, Peer reviewed by Dr. Abdolghader Pakniyat

At the bedside

Use HFNC for selected adult hypoxemic failure with close monitoring. Prefer NIV for acidotic COPD and CPAP/NIV for cardiogenic pulmonary edema.

Set flow for respiratory demand and FiO₂ for the prescribed saturation target. Reassess effort and ventilation as well as SpO₂.

Escalate promptly for an unsafe airway, refractory hypoxemia or exhaustion. Never wait for a ROX calculation.

Magdalena: the initial decision

Magdalena arrived with hypoxemia, restlessness, tachypnea and increased work of breathing. Her lungs were clear and she had no edema. She could not tolerate the oxygen mask and removed it. These are the reported case facts; her initial blood gases and hemodynamic measurements are not provided.

Mask intolerance alone does not determine the next device. Her airway safety, ventilation, perfusion and underlying diagnosis still need assessment.

Recognize deterioration before choosing a device

Breathlessness, tachypnea, hypoxemia and hypercapnia can appear in different combinations. They do not form a reliable stepwise sequence. Establish whether the main problem is oxygenation, ventilatory failure, or both. Saturation alone cannot assess carbon dioxide clearance.

Respiratory arrest, an unsafe airway or life-threatening hypoxemia despite maximal support requires immediate airway control and assisted ventilation. Worsening consciousness, rising oxygen need or shock demands urgent reassessment and an airway plan. SRLF/SFMU guidance, 2024.

What HFNC changes physiologically

HFNC supplies warmed, humidified gas through an open nasal interface. High flow better matches inspiratory demand and limits room-air dilution. Flow and FiO₂ are separate settings; delivered oxygen still depends on demand and interface performance. ERS HFNC guideline, 2022.

HFNC generates variable distending pressure influenced by flow, cannula fit, airway anatomy and mouth opening. It may increase end-expiratory lung volume, but provides neither prescribed PEEP nor dependable CPAP. Small physiological studies demonstrate this variability. Parke et al., 2009.

Upper-airway dead-space washout reduces rebreathing and can improve ventilatory efficiency. It does not guarantee alveolar ventilation, deliver a set tidal volume, or replace assisted ventilation during apnea. Möller et al., 2017.

A small crossover study found less inspiratory effort and greater end-expiratory lung volume than with a face mask. These mechanisms are plausible benefits, not proof of survival benefit. Mauri et al., 2017.

Conventional cannulae and masks provide variable inspired oxygen. Use device instructions rather than converting every L/min into a fixed FiO₂. A Venturi adapter specifies concentration and supply flow; reservoir masks need adequate flow to keep the bag inflated. BTS oxygen guideline, 2017.

What the adult trials show

The 2026 ATS guideline strongly recommends HFNC over standard oxygen for adults with de novo acute hypoxemic respiratory failure. It makes no recommendation favoring HFNC over NIV or CPAP over NIV. Selection still depends on the patient, interface expertise and available monitoring. ATS guideline, 2026.

FLORALI found no significant overall difference in its primary intubation outcome between HFNC, standard oxygen and NIV among 310 nonhypercapnic adults. A secondary 90-day mortality finding favored HFNC, but should not be generalized into a survival guarantee. FLORALI, 2015.

SOHO analyzed 1,110 adults with pulmonary infiltrates, PaO₂/FiO₂ ≤200 and respiratory rate >25/min. Mortality at day 28 was 14.6% in both groups. Intubation, a secondary outcome, occurred in 42.4% with HFNC versus 48.4% with standard oxygen. Serious events during spontaneous breathing occurred in 2.3% versus 1.1%. This establishes neither mortality benefit nor safety equivalence. SOHO, 2026.

RENOVATE compared HFNC with NIV across five groups, predominantly COVID-19 or nonimmunocompromised hypoxemia. A Bayesian model borrowed information across groups and allowed a noninferiority margin of odds ratio 1.55. Four groups met the criterion for intubation or death within seven days; immunocompromised hypoxemia did not. The COPD subgroup had only 77 patients and model-sensitive findings. This is not proof of superiority for pulmonary edema. RENOVATE, 2025.

The cases and the intubation pathway

The presentation reports that Magdalena received thrombolytics and was successfully supported with HFNC during recovery. It does not provide enough detail to identify the thrombolysis indication or attribute her improvement to HFNC alone.

Rajesh’s slide reports respiratory rate 30/min and accessory-muscle use. Its assumed tidal volume of 0.5 L would give 15 L/min minute ventilation, but the volume was not documented as measured. Respiratory rate alone cannot determine ventilation.

Preoxygenation is a separate task from ongoing HFNC treatment. ATS recommends HFNC or NIV over standard oxygen, without preferring one over the other. HFNC may continue during laryngoscopy, but apneic oxygenation does not provide dependable ventilation. ATS guideline, 2026.

PREOXI found peri-intubation hypoxemia below 85% in 9.1% with NIV preoxygenation versus 18.5% with an oxygen mask. It did not directly compare NIV with HFNC. Choose the approach around oxygenation, need for assisted breaths, aspiration risk and the airway plan. PREOXI, 2024.

Children and COVID-19: keep the populations separate

In Franklin’s trial of 1,472 infants under 12 months with bronchiolitis requiring oxygen outside ICU, escalation occurred in 12% with HFNC versus 23% with standard oxygen. This was not a 93% reduction in intubation and does not cover every pediatric diagnosis. Franklin et al., 2018.

Use a pediatric or neonatal pathway appropriate to the diagnosis, age, weight and equipment. A universal pediatric maximum of 50–60 L/min is inappropriate. Cannulae must allow the intended leak. Monitor effort, oxygen requirement, alertness, pressure injury and abdominal distension; reconsider pneumothorax with sudden deterioration. Oral feeding requires a separate assessment of stability and aspiration risk. RCH pediatric HFNP guidance.

Bronchiolitis also has different oxygen thresholds. NICE recommends supplemental oxygen for persistent SpO₂ below 90% from six weeks of age, or below 92% in younger infants and children with underlying conditions. These are treatment thresholds, not a universal HFNC saturation prescription. NICE bronchiolitis guidance.

RECOVERY-RS found no significant reduction in intubation or death with HFNO versus conventional oxygen in COVID-19. Early stopping and crossover limit interpretation; this does not establish equivalence. RECOVERY-RS, 2022.

Choose support and start it deliberately

Acidotic COPD: bilevel NIV usually takes priority because inspiratory pressure support unloads respiratory muscles. Acute cardiogenic pulmonary edema: CPAP or bilevel NIV has an established role when suitable. Controlled positive pressure improves mechanics and can reduce left-ventricular afterload; watch blood pressure and perfusion. Patients in cardiogenic shock were generally excluded from the supporting trials. ERS/ATS NIV guideline, 2017.

ATS retains a strong recommendation for NIV in acute hypercapnic failure. HFNC is conditional for selected less-severe cases with mild acidemia, for example pH >7.25, only with close observation and prompt NIV escalation. A pH threshold alone does not establish suitability. ATS guideline, 2026.

In Tan’s 225-patient COPD trial, with pH 7.25–7.35 and PaCO₂ ≥50 mmHg, HFNC failed its noninferiority test against NIV. Treatment failure, defined as intubation or switching support, was 25.7% versus 14.3%; intubation was 14.2% versus 5.4%. These findings support caution when substituting HFNC for NIV. Tan et al., 2024.

Acute asthma: these COPD-dominant findings do not establish benefit in asthma. Treat airflow obstruction, obtain expert help, and avoid delayed intubation in an exhausted or deteriorating patient. ERS/ATS NIV guideline, 2017.

Adult studies commonly use 40–60 L/min, adjusted to demand and tolerance. Use active humidification and the selected device’s cannula, circuit, mode and flow limits. Set FiO₂ separately to the prescribed oxygen target. ERS HFNC guideline, 2022.

Prescribe an oxygen target. BTS recommends SpO₂ 94–98% for most acutely ill adults and 88–92% for those at risk of hypercapnic failure pending blood gases and reassessment. In critical hypoxemia, provide sufficient oxygen immediately, then titrate promptly. A normal saturation on high FiO₂ can coexist with worsening hypercapnia. BTS oxygen guideline, 2017.

Reassess early and use ROX as an adjunct

Observe closely during initiation and after changes. Record effort, respiratory rate, SpO₂ with FiO₂, mental state, heart rate and blood pressure. Obtain blood gases when ventilation or acid-base status is uncertain. Agree who will escalate and where monitoring will continue. SRLF/SFMU guidance, 2024.

ROX = (SpO₂/FiO₂) ÷ respiratory rate. Enter SpO₂ as a percentage, FiO₂ as a fraction, and respiratory rate in breaths/min. Roca’s prospective validation studied adults with pneumonia receiving HFNC. ROX ≥4.88 at 2, 6 or 12 hours was associated with lower intubation risk; values <2.85, <3.47 and <3.85 at those respective times predicted higher failure risk. Intermediate values need continued reassessment. Roca et al., 2019.

The ROX trend complements examination; it does not measure airway protection, perfusion, PaCO₂ or pH. Accuracy was modest and improved over time in the original cohort. Those cutoffs are not universal across diagnoses, children or hypercapnic failure. Clinical deterioration overrides a reassuring value, and urgent escalation must never wait for the next timed calculation. Roca et al., 2019.

Know when the trial has failed

Persistent severe effort, exhaustion, worsening acidosis, hypoxemia or perfusion indicates failure despite apparent comfort. Address secretions and obstruction. Nasal trauma or severe obstruction may prevent effective cannula use. HFNC neither protects the airway nor treats apnea. SRLF/SFMU guidance, 2024.

An observational study associated intubation at least 48 hours after starting HFNC with higher mortality than earlier intubation. Confounding prevents a causal conclusion, and 48 hours is not a safe waiting period. The bedside concern is persisting with ineffective support as the patient deteriorates. Kang et al., 2015.

Check oxygen supply, humidification, cannula position and skin. Staff must be able to recognize failure and provide escalation promptly. If intubation is outside agreed goals, define the treatment ceiling and reassess comfort and burden.

A practical endpoint

Continue while effort, oxygen need and clinical condition improve. Wean by response and local protocol while treating the cause.

Adapted from Dr. Sabrina Berdouk’s September 2025 presentation at the Emergency Medicine, Trauma & Critical Care Updates conference in Abu Dhabi, UAE. Clinical references reviewed through October 2026.

Sources for bedside decisions

• ATS guideline, 2026
• ERS HFNC guideline, 2022
• ERS/ATS NIV guideline, 2017
• SOHO, 2026
• FLORALI, 2015
• RENOVATE, 2025
• Tan et al., 2024
• Roca et al., 2019
• Kang et al., 2015
• Parke et al., 2009
• Möller et al., 2017
• Mauri et al., 2017
• PREOXI, 2024
• Franklin et al., 2018
• RECOVERY-RS, 2022
• BTS oxygen guideline, 2017
• SRLF/SFMU guidance, 2024
• RCH pediatric HFNP guidance
• NICE bronchiolitis guidance