ICU · Respiratory / oxygen therapy
High-Flow Nasal Cannula — Mechanisms, Evidence & the ROX Index
Also known as High-flow nasal cannula · HFNC · HFNO · Optiflow · ROX index · FLORALI trial · Dead-space washout · Heated humidified high-flow
High-flow nasal cannula (HFNC) delivers up to 60 L/min of heated, humidified oxygen-enriched gas through large-bore nasal prongs. Its five physiological mechanisms are low-level PEEP, dead-space washout, heated humidification, reduced inspiratory resistance, and a precise titrated FiO2. The FLORALI trial (NEJM 2015) supports HFNC in acute hypoxaemic respiratory failure, with reduced intubation in pneumonia. The ROX index (SpO2/FiO2 divided by respiratory rate; Roca, J Crit Care 2016) predicts success and the need to intubate — at least 4.88 favours success, under 3.85 favours failure. HFNC is a bridge, not a destination: a failing patient must be intubated.
On this page & tools
Your progress
Saved locally on this device.
Target exams
Overview & definition
High-flow nasal cannula (HFNC) delivers up to 60 L/min of heated, humidified gas with a titrated FiO2 (up to 100 per cent) through large-bore nasal prongs. It sits between standard oxygen and non-invasive ventilation, and works through five physiological mechanisms rather than by a single effect. The FLORALI trial (NEJM 2015) supports its use in acute hypoxaemic respiratory failure, and the ROX index predicts which patients will succeed.[1][2]
The five physiological mechanisms
- Low-level PEEP — the high flow generates a small positive pressure (about 3-5 cmH2O at 30-60 L/min, higher with the mouth closed), which splints alveoli, recruits collapsed lung, and reduces work of breathing.[1]
- Dead-space washout — the high flow flushes the CO2-laden gas from the upper airway (the anatomic dead space) before each breath, so each tidal volume reaches the alveoli with fresh gas; this effectively raises alveolar ventilation and helps clear CO2.[1]
- Heated humidification — gas is delivered at about 37 degrees and fully humidified, which preserves mucociliary function, mobilises secretions, reduces their viscosity, and markedly improves comfort compared with dry oxygen.[1]
- Reduced inspiratory resistance — the delivered flow meets or exceeds the patient's peak inspiratory flow, so the patient does not have to draw gas through the nasal passages themselves; this unloads the inspiratory muscles and reduces the work of breathing.[1]
- A precise, titrated FiO2 — because the flow meets demand, the patient does not entrain room air, so the delivered FiO2 is accurate and controllable (unlike the low-flow variable-performance devices).[1]
Quantifying the mechanism — what the numbers actually are
The five mechanisms are not abstract: each is measurable, and the magnitude matters for the exam and the bedside. [1]
The five HFNC mechanisms — measurable effects at the bedside
| Mechanism | Measurable effect | Magnitude / what changes it | Clinical consequence |
|---|---|---|---|
| Low-level PEEP | Positive pressure in the pharynx (measured by oesophageal manometry) | ~1 cmH2O at 10 L/min, rising ~0.5-1 cmH2O per 10 L/min; ~3-5 cmH2O at 30-60 L/min. Mouth closed roughly doubles the pressure that reaches the alveoli vs mouth open | Alveolar recruitment, raised end-expiratory lung volume, reduced atelectrauma. Pressure is flow-dependent and leak-dependent — NOT a guaranteed number |
| Dead-space washout | Reduction in anatomic dead space (nasopharynx, oropharynx, trachea ~150 mL) | Effective dead space falls by ~25-50 per cent at 30-60 L/min. Each breath delivers more fresh gas to the alveoli for the same tidal volume | Higher alveolar ventilation at the same minute volume — PaCO2 typically falls by 2-5 mmHg even without ventilatory pressure support |
| Heated humidification | Absolute humidity delivered to the airway | ~33-44 mg H2O/L at 31-37 degrees (fully saturated). Dry wall oxygen delivers <10 mg H2O/L | Preserved mucociliary clearance, less inspissated secretion, less epithelial injury, far better tolerance. Humidification also reduces the metabolic cost of gas conditioning |
| Reduced inspiratory resistance | Inspiratory resistance of the nasopharynx largely bypassed | Delivered flow (30-60 L/min) meets or exceeds the patient's peak inspiratory demand (which can exceed 60 L/min in distress) | The inspiratory muscles do not have to pull gas through the nose; work of breathing and respiratory rate fall, often within minutes |
| Precise FiO2 | Delivered FiO2 matches the set value | At flows above ~30 L/min there is little room-air entrainment, so FiO2 is accurate from 0.21 to 1.0 | Allows true oxygen titration and a valid ROX/P-F calculation; permits lung-protective low-FiO2 strategy |
What HFNC actually does to the failing lung — the four measurable wins
- RAISES END-EXPIRATORY LUNG VOLUME — low-level PEEP splints alveoli that would otherwise collapse at end-expiration, increasing functional residual capacity and improving V/Q matching. Demonstrated on electrical impedance tomography within minutes of starting HFNC.[6]
- REDUCES RESPIRATORY RATE — by washing out dead space (each breath becomes more efficient) and unloading the inspiratory muscles, the patient slows down. A falling respiratory rate over the first hour is one of the earliest signs HFNC is working.
- REDUCES PaCO2 — dead-space washout improves alveolar ventilation without adding pressure support, so PaCO2 typically drops by a few mmHg. This is enough for mild-moderate hypercapnia but NOT for significant type-2 failure.
- IMPROVES OXYGENATION AND COMFORT — accurate FiO2, recruitment, and humidification combine. Patient-reported comfort is consistently higher than standard oxygen or a tight NIV mask.[6]
The evidence — FLORALI and the trial that changed practice
The FLORALI trial (Frat, NEJM 2015) was the pivotal study that moved HFNC from a comfort device to a first-line therapy for acute hypoxaemic respiratory failure. It is a non-blinded, multicentre randomised controlled trial comparing three strategies in patients with acute hypoxaemic respiratory failure (PaO2/FiO2 <=300 on standard oxygen for at least 15 minutes).[1]
FLORALI 2015 — High-flow oxygen in acute hypoxaemic respiratory failure (PMID 25981908)
Design
Multicentre, unblinded randomised controlled trial, 23 French ICUs, 310 patients with acute hypoxaemic respiratory failure (PaO2/FiO2 ratio <=300 on standard O2)
Three arms
HFNC (n=106) vs standard oxygen with non-rebreather mask (n=102) vs non-invasive ventilation with pressure support up to 10 cmH2O + PEEP 5-10 cmH2O delivered as intermittent one-hour sessions (n=102)
Primary outcome
28-day intubation rate — the trial's primary endpoint
Intubation rate (overall)
HFNC 38% vs standard O2 47% vs NIV 50% — NOT statistically significant overall (p=0.18), but a clear trend favouring HFNC
Pre-specified subgroups
In PNEUMONIA and in patients with PaO2/FiO2 <150, HFNC significantly REDUCED intubation vs standard oxygen. This subgroup signal is the clinically important finding
Secondary — 90-day mortality
HFNC 12% vs standard O2 23% vs NIV 28% — significantly LOWER with HFNC (hazard ratio 0.42, p=0.046 in favour of HFNC over the two controls combined)
Tolerance
HFNC better tolerated than NIV (mask discomfort, skin breakdown, asynchrony); median time on therapy longer with HFNC
Bottom line
In moderate acute hypoxaemic respiratory failure — especially pneumonia and PaO2/FiO2 <150 — HFNC is a legitimate FIRST-LINE therapy that REDUCES intubation and may reduce 90-day mortality. It does NOT replace intubation for severe or worsening failure. Established HFNC as standard of care
Reading FLORALI correctly — what it does and does not prove:
- The primary endpoint (28-day intubation) was negative overall — the headline is not "HFNC beats everything". The signal is in the pre-specified subgroups: pneumonia and PaO2/FiO2 <150, where HFNC clearly reduced intubation.
- 90-day mortality was lower with HFNC (12 per cent vs 23 per cent vs 28 per cent), a secondary endpoint that should be interpreted cautiously but is mechanistically plausible (fewer intubations = fewer ventilator-associated complications, less sedation, less delirium).
- NIV in this trial was delivered as intermittent sessions (median ~8 hours/day), which is the way NIV is often used for type-1 failure and is part of why NIV did poorly — current practice favours CPAP for cardiogenic oedema and continuous NIV in selected type-1 patients.[1]
The subsequent meta-analytic evidence — Ou 2017 (CMAJ)
| Outcome | HFNC vs standard oxygen | HFNC vs NIV |
|---|---|---|
| Intubation rate | Significantly LOWER with HFNC (RR ~0.65-0.7) | No significant difference |
| ICU mortality | Lower with HFNC (driven by FLORALI signal) | No significant difference |
| Escalation to NIV/intubation | Lower with HFNC | — |
| Bottom line | HFNC is at least as good as, and probably better than, standard oxygen for acute hypoxaemic respiratory failure | HFNC is non-inferior to NIV with better tolerance in type-1 failure |
The 2017 Ou meta-analysis pooled FLORALI with the other early RCTs and confirmed that HFNC reduces the need for escalation/intubation compared with standard oxygen in acute hypoxaemic respiratory failure, with no excess mortality.[7]
The ROX index
The ROX index (Roca, J Crit Care 2016) predicts the success of HFNC, derived and validated in pneumonia:[2]
ROX = (SpO2 / FiO2) divided by respiratory rate [1]
- A score at least 4.88 at 2, 6, and 12 hours predicts success (a low risk of intubation).
- A score under 3.85 predicts failure (a high risk of needing intubation).
- The threshold rises with time — HFNC that is working shows a stable or improving ROX. The validated cut-offs are time-specific (see table below) and apply to pneumonia; the trend matters as much as any single value.
- The trend matters — a falling ROX over time is more worrying than a single low value. Use it alongside the clinical picture (rising respiratory rate, increasing accessory-muscle use, exhaustion).[2]
ROX index thresholds by time on HFNC (Roca 2016, derived in pneumonia)
| Time after starting HFNC | ROX predicting SUCCESS (low intubation risk) | ROX predicting FAILURE (intubate) | Action if between the two |
|---|---|---|---|
| 2 hours | >= 4.88 | < 2.73 | Continue HFNC, monitor closely, repeat ROX |
| 6 hours | >= 4.88 | < 3.15 | Continue, but low and falling = prepare for intubation |
| 12 hours | >= 4.88 | < 3.85 | Low or falling ROX = intubate; do not delay |
| Trend | Stable or rising over time | Falling over time | A falling trend is the strongest warning — intubate |
Worked examples (ROX = (SpO2 % / FiO2 fraction) divided by respiratory rate).
- Success: SpO2 95%, FiO2 0.5, RR 25 → (95/0.5) = 190; 190/25 = 7.6 — above 4.88, on the right side of the line.
- Failure: SpO2 90%, FiO2 0.9, RR 35 → (90/0.9) = 100; 100/35 = 2.86 — below 3.85, the patient is failing; prepare to intubate. The index punishes a high FiO2 that buys only a borderline SpO2, and a high respiratory rate. A patient needing FiO2 0.9 and breathing 35 times a minute to hold SpO2 90% is failing HFNC even if the saturation looks acceptable. [1]
Indications
- Acute hypoxaemic respiratory failure (type 1) — pneumonia is the strongest evidence base; mild-moderate ARDS; post-operative and post-extubation respiratory failure.
- Preoxygenation before intubation, especially in the hypoxaemic patient (HFNC continues through the apnoeic period, extending safe apnoea time).
- Cardiogenic pulmonary oedema as a bridge (though CPAP/NIV has stronger evidence here).
- Do-not-intubate patients for symptom relief.[1]
Contraindications and cautions
- A patient who needs immediate intubation — arrest, obtundation, an unprotected airway.
- Type-2 (hypercapnic) failure — HFNC washes out dead space but does not provide the pressure support of bilevel NIV, so it is less effective than NIV for significant CO2 retention (use cautiously, escalate to NIV/intubation).[1]
- A facial or nasal deformity preventing a seal (a poor fit).
- An undrained pneumothorax (a relative caution with the PEEP effect).[1]
Practical setup and monitoring
- Flow: start at 30-40 L/min and titrate up to 60 L/min to the patient's comfort and work of breathing.
- FiO2: titrate to a target SpO2 of 92-96 per cent.
- Temperature: 31-37 degrees (lower temperature at higher flow for comfort, as cool gas feels less oppressive).
- Reassess at 1-2 hours with the ROX index and the clinical trend. Failure signs — a rising respiratory rate, increasing accessory-muscle use, a falling SpO2, altered mental status, haemodynamic instability — mean intubate now; do not let a failing HFNC trial delay a higher-risk intubation.[2][1]
Starting HFNC at the bedside — the first hour
- CHOOSE THE INTERFACE — select the correct nasal cannula size (prongs should fill about half the nares; too small wastes flow as leak). Connect to a blended, humidified HFNC system (e.g., Airvo, Optiflow, myAIRVO, Vapotherm).[1]
- SET FLOW FIRST, THEN FiO2 — begin at 30-40 L/min (lower in older/frail patients, e.g., 25-30 L/min, then titrate up). Set FiO2 to 1.0 initially to preoxygenate, then wean to the lowest FiO2 that holds SpO2 92-96 per cent (88-92 per cent in COPD/CO2 retainers).[1]
- SET TEMPERATURE 34-37 degrees — 37 degrees for maximal humidification (thick secretions, thick sputum); 31-34 degrees if the patient finds high flow claustrophobic (cooler gas feels less oppressive at high flow).
- REASSESS AT 15-30 MINUTES — respiratory rate, accessory-muscle use, SpO2, comfort, synchrony. A falling respiratory rate and relaxed accessory muscles within the first 30 minutes is an early sign of response.
- COMPUTE THE ROX INDEX at 1-2 hours and again at 6 and 12 hours. Plot the trend. A rising or stable ROX >= 4.88 supports continuation; a falling ROX or any value < 3.85 triggers the intubation conversation.
- TREAT THE CAUSE IN PARALLEL — antibiotics for pneumonia, diuretics for cardiogenic oedema, bronchodilators for COPD/asthma. HFNC is oxygenation support, not a treatment for the underlying disease.
- SET A CLEAR ESCALATION THRESHOLD before you walk away — define which deterioration triggers intubation (e.g., ROX < 3.85 at any point, RR > 35 sustained, SpO2 < 90% on FiO2 >= 0.9, exhaustion, altered mentation).[1][2]
HFNC vs NIV vs standard oxygen — choosing the device
The choice between standard oxygen, HFNC, and NIV depends on the type of respiratory failure (hypoxaemic vs hypercapnic), the expected trajectory, the need for pressure support, and tolerance. No device is universally superior — match the device to the pathophysiology. [1]
HFNC vs NIV vs standard oxygen — head-to-head
| Feature | Standard oxygen (NRB mask, nasal specs) | HFNC | NIV (bilevel/CPAP) |
|---|---|---|---|
| Max flow | ~15 L/min (NRB) | 30-60 L/min | High (driven by turbine) |
| FiO2 delivered | Variable (room-air entrainment) — 0.4-0.9 in practice | Accurate, 0.21-1.0 | Accurate, 0.21-1.0 |
| Pressure support | None | Low-level PEEP ~1-5 cmH2O (flow/mouth dependent) | True IPAP/EPAP, typically 8-20/5-10 cmH2O |
| Dead-space washout | Minimal | Significant — reduces effective dead space 25-50% | Moderate (mask adds dead space; pressure does the work) |
| Humidification | Poor (dry) | Excellent (37 degrees, fully humidified) | Variable (depends on circuit humidifier) |
| Best for | Mild hypoxaemia, stable patients, ambulatory oxygen | Acute hypoxaemic (type-1) failure — pneumonia, mild-moderate ARDS; preoxygenation; post-extubation | Hypercapnic (type-2) failure — COPD, neuromuscular; cardiogenic pulmonary oedema (CPAP) |
| Worst for | Anything more than mild hypoxaemia | Significant hypercapnia (use NIV instead) | Ventilatory failure where patient cannot protect airway; claustrophobic or asynchronic patients |
| Tolerance | Best | Excellent (no mask, patient can talk, eat) | Often poor — mask discomfort, skin breakdown, claustrophobia, asynchrony |
| Evidence (type-1 failure) | Standard-of-care comparator | FLORALI — reduced intubation in pneumonia subgroup | Mixed; intermittent NIV in type-1 failure did not beat HFNC in FLORALI |
| Can eat/talk | Yes | Yes | Difficult (mask) |
HFNC for preoxygenation before intubation
Endotracheal intubation in the hypoxaemic patient is dangerous — desaturation during the apnoeic period is rapid and is associated with cardiac arrest and death. HFNC is the best preoxygenation device available because it provides apnoeic oxygenation: the high flow continues to deliver oxygen to the alveoli throughout the apnoeic period of laryngoscopy, extending safe apnoea time. [1]
- Set up HFNC at 50-60 L/min, FiO2 1.0, for at least 3-5 minutes of tidal-volume preoxygenation before induction (denitrogenation).
- Leave the HFNC nasal cannula IN PLACE during laryngoscopy — the high flow continues to wash CO2 from the pharynx and oxygenate the alveoli (apnoeic oxygenation via the pressure gradient between pharynx and alveoli). This is "transnasal humidified rapid-insufflation ventilatory exchange" (THRIVE) physiology applied to the intubating patient.
- Caution: HFNC preoxygenation does NOT protect against desaturation in the patient with severe shunt (dense consolidation, ARDS) — these patients desaturate regardless because of venous admixture. Plan for rapid intubation and have a back-up plan.
- The randomised evidence is mixed — some trials (Semler 2016, Vourc'h) show no reduction in severe desaturation with HFNC vs NRB mask, but HFNC at least does no harm and provides the best tolerated high-FIO2 preoxygenation in the awake, distressed patient.[1]
Post-extubation HFNC — prevention and rescue
Post-extubation respiratory failure (reintubation within 48-72 hours) occurs in 10-20 per cent of patients and carries a high mortality. Two Hernández trials (JAMA 2016) defined the role of HFNC after extubation, and they are commonly examined. [1]
Hernández 2016 (low-risk) — Post-extubation HFNC vs conventional oxygen (PMID 26975498)
Design
Multicentre randomised controlled trial, 527 mechanically ventilated patients at LOW risk for reintubation (>= 65 years OR cardiac/respiratory failure as cause of ventilation)
Intervention
HFNC at 30 L/min for 24 h after extubation vs conventional oxygen therapy (NRB mask / nasal specs)
Primary outcome
Reintubation within 7 days: 4.9% (HFNC) vs 12.2% (conventional), p=0.004 — a 60% relative risk reduction
Secondary
Lower post-extubation respiratory failure, fewer ICU readmissions
Bottom line
In low-risk (or older) patients, prophylactic HFNC after extubation halves reintubation compared with conventional oxygen. Consider it as the default post-extubation support in older or cardiorespiratory patients
Hernández 2016 (high-risk) — Post-extubation HFNC vs NIV (PMID 27706464)
Design
Multicentre randomised controlled trial, 604 patients at HIGH risk for reintubation (>= 65 yrs with cardiac/respiratory cause, obesity-hypoventilation, COPD, prolonged ventilation, multiple failed weaning)
Intervention
HFNC for 24 h vs prophylactic NIV (bilevel, intermittent sessions) after extubation
Primary outcome
Reintubation within 7 days: 22.8% (HFNC) vs 19.1% (NIV) — NON-INFERIORITY of HFNC was NOT met (the upper bound exceeded the non-inferiority margin)
Secondary — mortality
ICU mortality LOWER with HFNC: 1.6% vs 6.8% (p=0.01), driven partly by fewer cardiac-related post-extubation deaths
Bottom line
In high-risk patients, NIV remains the default post-extubation support (HFNC failed non-inferiority for reintubation). However HFNC reduced ICU mortality — likely because NIV-induced swings delayed reintubation in failures. Use NIV first in high-risk, but HFNC is a reasonable alternative where NIV is refused or not tolerated
Reading the two Hernández trials together:
- Low-risk/older → HFNC is the post-extubation support of choice (better than conventional oxygen).
- High-risk → NIV remains first-line, but HFNC is a viable alternative in the NIV-intolerant or NIV-contraindicated patient, and may reduce ICU mortality.[4][5]
Post-operative HFNC — cardiothoracic and major surgery
After cardiothoracic and major abdominal surgery, hypoxaemia from atelectasis, splinting, and pulmonary oedema is common. The Stéphan FLAIR trial (JAMA 2015) showed HFNC is non-inferior to NIV for post-operative hypoxaemic respiratory failure after cardiothoracic surgery, with better tolerance.[3]
Stéphan 2015 (FLAIR) — HFNC vs NIV after cardiothoracic surgery (PMID 25980660)
Design
Multicentre randomised non-inferiority trial, 820 patients with hypoxaemic respiratory failure (PaO2/FiO2 <=300) within 7 days of cardiothoracic surgery
Intervention
HFNC vs NIV (bilevel, via face mask, intermittent sessions)
Primary outcome
Treatment failure (reintubation, switch therapy, or premature discontinuation) by day 7: non-inferiority met — 22% HFNC vs 22% NIV
Secondary
Better tolerance with HFNC; fewer skin breakdown events; median 15-minute sessions of NIV were not tolerated in ~40% of the NIV group
Bottom line
In post-cardiothoracic-surgery hypoxaemic respiratory failure, HFNC is non-inferior to NIV and better tolerated — a reasonable first-line choice. NIV reserved for failure or hypercapnia
This trial extended HFNC from medical AHRF into the surgical recovery setting, where mask intolerance and the need to communicate make HFNC particularly attractive. [1]
Special populations and contexts
HFNC in specific clinical contexts — what the evidence supports
| Context | Role of HFNC | Evidence / comment |
|---|---|---|
| Pneumonia (CAP/HAP) | First-line for hypoxaemic failure | FLORALI subgroup — reduced intubation; ROX index validated here. The strongest single indication.[1][2] |
| Mild-moderate ARDS | Acceptable alternative to NIV; bridge to intubation/ECMO | Avoid HFNC in SEVERE ARDS (P/F <150 with high work of breathing) — these patients usually need intubation for lung-protective ventilation |
| COVID-19 / viral pneumonitis | Widely used as 'first-line' oxygen; pause to assess at 1 h | High flows of 60 L/min generate aerosol — use airborne precautions. ROX applies. Awake proning complements HFNC |
| Cardiogenic pulmonary oedema | Adjunct if NIV/CPAP not tolerated | CPAP/NIV has stronger evidence (3CPO, L'Her). HFNC is reasonable if patient refuses a mask |
| COPD with mild hypercapnia | Can be used with close CO2 monitoring | For significant hypercapnia (pH < 7.30, PaCO2 > 70), NIV is first-line. HFNC acceptable if PaCO2 normal/mildly raised and patient NIV-intolerant |
| Immunocompromised | Increasing evidence as first-line | Avoids intubation (a high-mortality event in this group). Some trials support; definitive SOHO-type data are evolving |
| Do-not-intubate (DNI) / palliative | Excellent for symptom relief | Provides humidified high-flow oxygen without a mask; keeps the patient comfortable and able to communicate |
| Post-extubation (low risk / older) | First-line preventive support | Hernández 2016 — halved reintubation vs conventional O2[5] |
| Preoxygenation before intubation | Best available device; leave running through apnoea | Provides apnoeic oxygenation. Severe shunt will still desaturate |
Weaning and stepping down from HFNC
HFNC is started as a bridge; the goal is to step down to standard oxygen once the underlying process improves. Wean when the patient is clinically improving and the oxygen requirement is falling. [1]
Stepping down from HFNC to standard oxygen
- CONFIRM RECOVERY — respiratory rate < 25, no accessory-muscle use, comfortable, improving radiographs/cause treated, haemodynamically stable.
- TITRATE FiO2 DOWN FIRST to 0.3-0.4 while keeping the flow up; ensure SpO2 stays 92-96 per cent. Do not rush — a falling FiO2 requirement is the most reliable sign of recovery.
- REDUCE FLOW IN STEPS — 60 → 50 → 40 → 30 L/min over hours to a day, reassessing at each step. Watch the respiratory rate closely: a rise of > 5/minute after a step-down is a sign to hold or go back up.
- CONVERT TO STANDARD OXYGEN (NRB or nasal cannulae at an equivalent FiO2) once flow is tolerated at 30 L/min and FiO2 <= 0.35-0.4 with a stable respiratory rate. Use a slightly higher FiO2 than calculated to provide a safety margin.
- MONITOR FOR 1-2 HOURS after the switch for rebound work of breathing or desaturation. Rebound hypoxaemia in the first hour is common — if the respiratory rate climbs, return to HFNC.[1]
Adverse effects, troubleshooting, and pitfalls
HFNC is generally safe and well tolerated, but complications and pitfalls exist:
- Nasal/nare skin breakdown and pressure injury from the prongs — protect the skin with a hydrocolloid dressing; ensure prongs are not too large.
- Gastric insufflation is uncommon (much less than with NIV) but can occur at very high flows or in obtunded patients — place a nasogastric tube if needed.
- Mucosal dryness if humidification fails — check the water chamber and temperature; dry gas defeats the humidification mechanism.
- Aerosol generation / infection control — HFNC at 60 L/min generates some aerosol; use airborne/contact precautions for suspected respiratory pathogens (especially COVID-19, influenza, TB) until cleared.
- Misleading SpO2 / masked deterioration — the cardinal pitfall. SpO2 is held high by the FiO2 while work of breathing climbs silently. Watch the respiratory rate, accessory muscles, and ROX trend.[1][2]
- Delaying inevitable intubation — the most dangerous failure mode. A patient destined to fail HFNC who stays on it for hours has a more difficult, higher-risk intubation than one intubated early.
- Inadequate flow — using 20-25 L/min as a compromise rarely delivers the mechanism; the dead-space washout and PEEP effects are flow-dependent. Commit to 30 L/min minimum if you want a true HFNC trial.
Other prediction scores and clinical assessment
The ROX index is the best-validated HFNC score, but it is not the only tool. The HACOR score (Heart rate, Acidosis, Consciousness, Oxygenation, Respiratory rate) was designed for NIV and is sometimes applied to HFNC — a HACOR > 5 at 1 hour predicts NIV/HFNC failure. Beyond scores, serial clinical assessment (respiratory rate trend, accessory-muscle use, mental status, urine output, lactate) remains the cornerstone: no score replaces the bedside observation that the patient is tiring. [1]
Scores and signs predicting HFNC failure
| Tool | What it measures | Cut-off / interpretation | Comment |
|---|---|---|---|
| ROX index | (SpO2/FiO2) / RR | >= 4.88 success; < 3.85 failure (time-dependent) | Best validated; derived in pneumonia[2] |
| HACOR | Heart rate, acidosis (pH), consciousness, oxygenation, RR (for NIV) | > 5 at 1 h predicts failure | Originally for NIV; extrapolated to HFNC |
| Clinical trend | RR, accessory muscles, mental status, SpO2, lactate | Rising RR, exhaustion, falling SpO2 = failing | The most important and universally applicable |
| Serial ABG / VBG | pH, PaCO2, PaO2 | Falling pH, rising PaCO2/lactate = failing | Especially important in hypercapnic or mixed failure |
HFNC is a bridge, not a destination
The most important clinical principle: HFNC can mask deterioration by keeping the SpO2 acceptable while the work of breathing silently climbs. Watch the respiratory rate, the accessory-muscle use, and the ROX trend, not just the SpO2. A patient who is failing HFNC must be intubated — the longer they stay on a failing therapy, the higher the intubation risk.[1][2]
[1]SAQ — HFNC mechanisms and the FLORALI trial
10 minutes · 10 marks
A 62-year-old man with community-acquired pneumonia is hypoxaemic (SpO2 90% on a non-rebreather at 15 L/min, RR 30, PaO2 62 on FiO2 ~0.9). The registrar asks whether to start high-flow nasal cannula (HFNC) or proceed to non-invasive ventilation or intubation. The examiners ask you to explain how HFNC works, summarise the FLORALI trial, and justify your choice.
SAQ — The ROX index and detecting HFNC failure
10 minutes · 10 marks
A 71-year-old woman with pneumonia started HFNC (50 L/min, FiO2 0.6) two hours ago. Her SpO2 is now 94%, RR 28, and the registrar asks if HFNC is 'working.' The examiners ask you to define and apply the ROX index, describe how you would monitor this patient, and state the threshold for intubation.
Red flags
The landmark HFNC trials — what each changed
The key HFNC trials — design, result, and what it changed in practice
| Trial (year) | Population | Comparison | Key result | What it changed |
|---|---|---|---|---|
| FLORALI (2015)[1] | AHRF (PaO2/FiO2 <=300), n=310 | HFNC vs standard O2 vs NIV (intermittent) | 28-day intubation negative overall; subgroup-positive in pneumonia and P/F <150; 90-day mortality lower with HFNC | HFNC = first-line for moderate AHRF, especially pneumonia |
| Stéphan FLAIR (2015)[3] | Post-cardiothoracic surgery AHRF, n=820 | HFNC vs NIV (intermittent) | Non-inferior for treatment failure; better tolerance | HFNC first-line for post-cardiothoracic AHRF |
| Hernández low-risk (2016)[5] | Low-risk post-extubation (>=65 yr or cardiac/resp), n=527 | HFNC vs conventional O2 | Reintubation 4.9% vs 12.2% — halved | Prophylactic HFNC after extubation in older/low-risk patients |
| Hernández high-risk (2016)[4] | High-risk post-extubation, n=604 | HFNC vs NIV (prophylactic) | HFNC failed non-inferiority for reintubation; ICU mortality lower (1.6% vs 6.8%) | NIV first-line in high-risk; HFNC if NIV not tolerated |
| Roca ROX (2016)[2] | Pneumonia on HFNC, n=191 (derivation + validation) | Derivation of ROX index | ROX >= 4.88 predicts success; < 3.85 predicts failure (time-dependent) | The bedside prediction score for HFNC success |
| Mauri physiology (2017)[6] | AHRF on HFNC, physiological study | Measured pressures, EELV, work of breathing | Demonstrated measurable PEEP, recruitment, reduced RR/WOB | Quantified the five mechanisms; supports flow-dependent PEEP |
| Ou meta-analysis (2017)[7] | Adults with AHRF, pooled RCTs | HFNC vs standard O2 vs NIV | HFNC reduced intubation vs standard O2; non-inferior to NIV | Confirmed FLORALI signal across the early RCT evidence base |
Bottom line
High-flow nasal cannula is a five-mechanism respiratory support device — low-level PEEP, dead-space washout, heated humidification, reduced inspiratory resistance, and a precise titrated FiO2 — that sits between standard oxygen and non-invasive ventilation. FLORALI (NEJM 2015) established it as first-line for moderate acute hypoxaemic respiratory failure, especially pneumonia and PaO2/FiO2 <150, where it reduced intubation and may reduce 90-day mortality.[1] The ROX index (Roca 2016) — (SpO2/FiO2) divided by respiratory rate, with time-specific thresholds (>= 4.88 success, < 3.85 failure) — is the bedside tool to predict success and to decide when to intubate.[2] For type-2 (hypercapnic) failure, NIV remains first-line; HFNC is the device of choice for post-extubation support in low-risk/older patients (Hernández 2016) and is non-inferior to NIV after cardiothoracic surgery (Stéphan 2015).[3][4][5] The cardinal principle — repeated because it is the commonest source of harm — is that HFNC is a bridge, not a destination: it can mask deterioration by holding the SpO2 high while the work of breathing climbs silently. Watch the respiratory rate, the accessory-muscle use, the ROX trend, and the clinical trajectory; intubate the failing patient early.
References
- [1]Frat JP, Thille AW, Mercat A, et al.; FLORALI Study Group; REVA Network. High-flow oxygen through nasal cannula in acute hypoxemic respiratory failure N Engl J Med, 2015.PMID 25981908
- [2]Roca O, Messika J, Caralt B, et al. Predicting success of high-flow nasal cannula in pneumonia patients with hypoxemic respiratory failure: The utility of the ROX index J Crit Care, 2016.PMID 27481760
- [3]Stéphan F, Barrucand B, Petit P, et al. High-Flow Nasal Oxygen vs Noninvasive Positive Airway Pressure in Hypoxemic Patients After Cardiothoracic Surgery: A Randomized Clinical Trial JAMA, 2015.PMID 25980660
- [4]Hernández G, Vaquero C, Colinas L, et al. Effect of Postextubation High-Flow Nasal Cannula vs Noninvasive Ventilation on Reintubation and Postextubation Respiratory Failure in High-Risk Patients: A Randomized Clinical Trial JAMA, 2016.PMID 27706464
- [5]Hernández G, Vaquero C, González P, et al. Effect of Postextubation High-Flow Nasal Cannula vs Conventional Oxygen Therapy on Reintubation in Low-Risk Patients: A Randomized Clinical Trial JAMA, 2016.PMID 26975498
- [6]Mauri T, Turrini C, Eronia N, et al. Physiologic Effects of High-Flow Nasal Cannula in Acute Hypoxemic Respiratory Failure Am J Respir Crit Care Med, 2017.PMID 27997805
- [7]Ou X, Hua Y, Liu J, Gong C, Zhao W. Effect of high-flow nasal cannula oxygen therapy in adults with acute hypoxemic respiratory failure: a meta-analysis of randomized controlled trials CMAJ, 2017.PMID 28246239