noninvasive respiratory support

in review 4 min read Updated 2026-10-29
noninvasive respiratory support

Hypoxaemic failure → HFNC (over conventional O₂ and over NIV). Hypercapnic acidotic COPD (pH ≤7.35) → BiPAP (strong, high certainty; NIV first, HFNC only if NIV fails or isn’t tolerated). Cardiogenic pulmonary oedema → CPAP or BiPAP (strong). After extubation: low/moderate risk → HFNC; high risk → prophylactic NIV. Don’t use NIV for established post-extubation failure or for non-acidotic hypercapnia. Set a time limit: reassess at 1–2 h, and intubate early rather than late when it isn’t working.


which device for which patient

scenariofirst choicestrengthnotes
de novo hypoxaemic failure (pneumonia, early ARDS)HFNCERS conditional (vs COT moderate; vs NIV very low certainty); ESICM strong vs COT (less intubation); SSC 2026 applies it to P/F <200; ESICM no recommendation HFNC vs NIVno mortality difference; fewer intubations. ERS/ATS 2017 makes no NIV recommendation for de novo failure
COPD, pH ≤7.35, PaCO₂ >45, RR >20–24BiPAPstrong, highno pH floor, but the lower the pH the closer to intubation. Also before intubation unless crashing
COPD hypercapnic, not acidoticno NIVconditional againsttreat the exacerbation, controlled O₂ 88–92%
cardiogenic pulmonary oedemaCPAP or BiPAPstrongmortality RR 0.80; not in cardiogenic shock → acute decompensated heart failure
immunocompromisedHFNC or early NIVNIV conditional (2017)hypoxaemic, so the HFNC-over-NIV recommendation (2022) applies too
post-operative ARF (abdominal, thoracic, cardiac)NIV or HFNCconditionalsource control first (anastomotic leak, collection). Upper GI surgery isn’t a contraindication if the surgeon agrees
post-extubation, low/moderate riskHFNCconditionalover COT; prophylactic NIV not suggested
post-operative, low riskCOT or HFNCconditional
post-extubation, high risk (age >65, cardiac or respiratory disease, hypercapnia)NIV, prophylacticERS/ATS conditional; CHEST/ATS 2017 strongHFNC if NIV contraindicated; HFNC between NIV sessions
established post-extubation failurereintubateconditional against NIVNIV here delayed reintubation and increased mortality
weaning hypercapnic patientsextubate to NIVconditional
chest traumaNIV (pain controlled)conditional
asthmacautious trial onlyno recommendation→ acute asthma
palliative dyspnoea (terminal cancer)NIV if it helps the symptomconditionallowers opioid needs; not if it prolongs dying against goals
non-intubated hypoxaemic failure (sepsis, COVID-19)trial of awake proningSSC 2026 conditional, very low; ESICM suggests in COVID-19
COVID-19 hypoxaemiaCPAP/NIV or HFNCESICM weak (CPAP less intubation)

HFNC

  • heated, humidified gas at up to 60 L/min with FiO₂ up to 1.0. High flow meets peak inspiratory demand, so FiO₂ is what’s set. It washes out nasopharyngeal dead space (↓ PaCO₂ slightly), adds a little PEEP (~2–5 cmH₂O, mouth closed) and is comfortable, so it’s tolerated for days
  • start at 50–60 L/min and FiO₂ to SpO₂ 92–96% (88–92% if hypercapnia risk); wean FiO₂ first, then flow
  • FLORALI (2015): HFNC vs O₂ vs NIV in non-hypercapnic hypoxaemic failure. Intubation was similar overall, lower with HFNC when P/F ≤200, and 90-day mortality was lower with HFNC than with either. The likely reason NIV loses: large tidal volumes on NIV injure the lung (patient self-inflicted lung injury)
  • ROX index = (SpO₂/FiO₂) ÷ RR. ≥4.88 at 2, 6 or 12 h predicts HFNC success; <3.85 at 12 h → intubate (Roca, AJRCCM. 2019)
  • aerosol-generating: airborne precautions in suspected airborne infection
  • use HFNC between NIV sessions (meals, mask breaks)

NIV (CPAP and BiPAP)

  • CPAP: one pressure. Recruits alveoli and cuts LV afterload and preload, which is ideal in cardiogenic pulmonary oedema. Adds no ventilation
  • BiPAP: EPAP (= PEEP) + IPAP. The gap (pressure support) carries ventilation and clears CO₂. Typical start is IPAP 10–12 / EPAP 4–5 cmH₂O ; raise IPAP in steps of 2 to target RR and pH, and raise EPAP for oxygenation or to splint the upper airway in obstructive sleep apnoea. Oronasal mask
  • COPD logic: tired muscles under dynamic hyperinflation. Pressure support offloads them, and EPAP offsets intrinsic PEEP, reducing the effort needed to trigger each breath

Contraindications (most are relative): cardiac or respiratory arrest; can’t protect the airway or clear secretions; vomiting or upper GI bleed; facial trauma, burns or surgery; fixed upper airway obstruction; haemodynamic instability; untreated pneumothorax; uncooperative or agitated. A drowsy patient with hypercapnic encephalopathy is the exception: a closely monitored trial is reasonable, because CO₂ clearance often wakes them.

failure: when to stop and intubate

  • reassess at 1–2 h: pH and PaCO₂ not improving (COPD), RR still high, accessory muscles, rising FiO₂, falling consciousness, can’t tolerate the mask, haemodynamic instability
  • hypoxaemic failure: large expired tidal volumes on NIV predict failure, and a delayed intubation worsens outcome. Set the decision point up front
  • decide the ceiling of care (NIV as ceiling vs full escalation) at the start → goals of care

Intubation and invasive ventilation: mechanical ventilation. ARDS on HFNC or NIV meets the global definition: ards.

Prepared 2026-10-29. Not yet verified — check every number against the sources.

Key references

+3 more sources

Suggest a correction noninvasive respiratory support