Acute respiratory failure: classification and approach · ICU Topics

Cinematic ICU scene of a breathless patient on high-flow nasal cannula with a pulse oximeter reading low, an arterial blood gas syringe and a chest X-ray on screen, a ventilator standing by, clinical-blue lighting, medical educational, no faces, no text — FigureAcute respiratory failure is classified by the mechanism — type 1 (hypoxaemic, PaO2 <8 kPa) from V/Q mismatch, shunt or low inspired oxygen, and type 2 (hypercapnic, PaCO2 >6 kPa) from alveolar hypoventilation or ventilatory failure. The key physiological divider is the A-a gradient: normal gradient points to hypoventilation (opioid, neuromuscular); a widened gradient points to lung disease. Escalate from oxygen to high-flow nasal cannula to non-invasive ventilation to intubation along the failure trajectory.

[1] [1]

Type 1 vs Type 2

Type 1 (Hypoxaemic)

PaO2 <60 mmHg, normal/low PaCO2

A-a gradient ELEVATED (lung problem — gas exchange impaired)
Mechanism: V/Q mismatch (#1), shunt, diffusion impairment
Causes: pneumonia, ARDS, PE, pulmonary oedema, asthma, interstitial lung disease
Treatment: oxygen therapy (high FiO2 for V/Q mismatch, PEEP for shunt)
Ventilation: if severe — NIV or intubation with lung-protective ventilation

Type 2 (Hypercapnic)

PaCO2 >45 mmHg

A-a gradient NORMAL (ventilation problem — alveolar hypoventilation)
Mechanism: reduced alveolar ventilation OR increased CO2 production OR V/Q mismatch with fatigue
Causes: COPD (#1), neuromuscular disease (MG, GBS, ALS), CNS depression (opioids, stroke), obesity hypoventilation, kyphoscoliosis
Treatment: ventilatory support (increase minute ventilation) — NIV (BiPAP) first-line for COPD
Oxygen: controlled (88-92% in COPD — avoid hyperoxia which worsens hypercapnia via Haldane effect)

[1] [2]

A-a gradient

[2]

Causes of hypoxaemia

5 mechanisms of hypoxaemia

V/Q mismatch (most common)

Ventilation-perfusion mismatch: some alveoli are perfused but not ventilated (pneumonia, atelectasis) or ventilated but not perfused (PE). Net effect: blood leaving lungs has less oxygen. RESPONDS to oxygen therapy (increased FiO2 corrects low V/Q units). Causes: pneumonia, COPD, asthma, pulmonary oedema.

Shunt

Blood bypasses ventilated alveoli completely → mixes with oxygenated blood → reduces overall PaO2. Does NOT respond to oxygen therapy (shunted blood never reaches alveoli). Causes: ARDS, lobar pneumonia, pulmonary AV malformation, intracardiac shunt (right-to-left). Treatment: PEEP (recruits collapsed alveoli), treating underlying cause.

Hypoventilation

Reduced alveolar ventilation → less fresh gas reaches alveoli → PaO2 falls AND PaCO2 rises. A-a gradient is NORMAL (lungs are fine — the pump is failing). Responds to oxygen AND to restoring ventilation. Causes: CNS depression (opioids, stroke), neuromuscular disease (MG, GBS), chest wall disorders.

Diffusion impairment

Thickened alveolar-capillary membrane → slower gas diffusion → less oxygen reaches blood. Rare acute cause. Causes: interstitial lung disease, pulmonary fibrosis. Responds to oxygen therapy.

Low inspired oxygen

High altitude, low FiO2 delivery, anaesthesia circuit disconnect. Corrects by increasing FiO2.

[2]

Clinical pearls

High-yight respiratory failure points for the CICM/FFICM exam

Type 1: PaO2 <60, normal CO2, high A-a gradient → lung disease.[1] }
Type 2: PaCO2 >45, normal A-a gradient → pump failure (CNS, neuromuscular, chest wall).[1] }
A-a gradient: distinguishes lung disease (high) from hypoventilation (normal).[2] }
V/Q mismatch: MOST COMMON cause, RESPONDS to oxygen therapy.[2] }
Shunt: does NOT respond to oxygen → needs PEEP.[2] }
Haldane effect: hyperoxia worsens hypercapnia in COPD (O2 displaces CO2 from Hb).[2] }
Oxygen targets: Type 1 (92-96%), COPD/Type 2 (88-92%).[1] }
NIV: first-line for COPD Type 2 respiratory failure (BiPAP).[1] }
Fatigue = impending arrest: paradoxical breathing, inability to speak full sentences, accessory muscle use → prepare to intubate.[1] }
Lung-protective ventilation: VT 6 mL/kg PBW, plateau <30, driving pressure <15.[1] }
Clinical signs of respiratory distress: RR >30, SpO2 <90%, cyanosis, accessory muscle use, paradoxical breathing, inability to speak, agitation, altered mental status.[1] }
Venous blood gas: can estimate PaCO2 ( PvCO2 ≈ PaCO2 + 4-6 mmHg). Less invasive for monitoring CO2.[1] }
pH <7.30 with high CO2: decompensated Type 2 → ventilatory support needed.[1] }
Oxygen dissociation curve: steep below PaO2 60 mmHg → small drop in PaO2 = large drop in SaO2.[2] }

Red flags

Oxygen-haemoglobin dissociation curve

The oxyhaemoglobin dissociation curve — why SaO2 lies to you

The relationship between PaO2 and SaO2 is sigmoid (S-shaped), not linear. This has profound clinical consequences. [1]

Key landmarks:

P50 = 27 mmHg (PaO2 at which haemoglobin is 50% saturated)
PaO2 60 mmHg → SaO2 ~90% (the "shoulder")
PaO2 100 mmHg → SaO2 ~97-98%
PaO2 >100 mmHg → SaO2 barely changes (the flat upper portion) [1]

The flat upper portion (PaO2 >60): large changes in PaO2 cause only small changes in SaO2. A fall in PaO2 from 100 to 70 may not change the pulse oximeter reading at all — SpO2 stays ~95%. Pulse oximetry is a LATE warning sign of deterioration in this range. [1]

The steep lower portion (PaO2 <60): small falls in PaO2 cause large falls in SaO2. Below the shoulder, a drop from 60 to 40 mmHg plummets SaO2 from 90% to 75%. This is why a patient crossing SpO2 90% is on the edge of a cliff. [1]

Right shift (decreased affinity, more O2 release to tissues) caused by:

Increased H+ (acidosis), increased PaCO2 (Bohr effect)
Increased temperature, increased 2,3-DPG (chronic hypoxaemia, anaemia, high altitude) [1]

Left shift (increased affinity, less O2 release) caused by:

Alkalosis, hypocapnia, hypothermia, decreased 2,3-DPG (stored bank blood), CO poisoning (carboxyhaemoglobin), methaemoglobinaemia [1]

Clinical implications:

SpO2 92-96% corresponds to PaO2 60-100 mmHg — a SAFE zone with reserve.
SpO2 <90% = PaO2 <60 mmHg — steep part of curve, decompensation imminent.
Do not chase SpO2 100% — adds minimal O2 content but exposes patient to FiO2 toxicity.
CO poisoning: SpO2 is FALSELY normal (pulse oximeter cannot distinguish HbO2 from HbCO). Measure CO-Hb on ABG.
Each gram of Hb carries 1.34 mL O2. O2 content = (1.34 × Hb × SaO2) + (0.003 × PaO2). Dissolved O2 (the PaO2 term) is negligible at normal FiO2 — most O2 is bound to Hb. Anaemia causes tissue hypoxia WITHOUT low SaO2.

[2]

Bohr effect vs Haldane effect — high-yield exam distinction

Bohr effect (at the tissues): increased CO2 / H+ shifts the O2-Hb curve RIGHT → enhances O2 unloading where metabolism is high. A self-regulating mechanism.
Haldane effect (at the lungs): oxygenation of Hb in the lungs reduces its affinity for CO2 → promotes CO2 release into alveoli for exhalation. [1]

Clinical trap in COPD: Giving high-flow O2 to a CO2-retainer does TWO harmful things:

Haldane effect: oxygenated Hb releases CO2 → increased CO2 load to be exhaled, but the patient cannot increase ventilation → PaCO2 rises.
V/Q mismatch worsening: O2 abolishes hypoxic pulmonary vasoconstriction in poorly ventilated units → increased perfusion to low-V/Q lung → increased dead space and worsened hypercapnia. [1]

This is why controlled O2 (target SpO2 88-92%) is mandatory in COPD at risk of hypercapnia.

[1]

Work of breathing (WOB)

Work of breathing is the oxygen cost the respiratory muscles expend to move gas in and out. Normally ~2-5% of total O2 consumption, but in respiratory failure can rise to >30% — the muscle pump fatigues and fails.[2]

Components of WOB

Elastic work

~65% of total WOB

Work done to overcome the elastic recoil of lung and chest wall
Increased by: low lung compliance (ARDS, pulmonary oedema, fibrosis, atelectasis, pneumothorax)
Reduced compliance means more pressure (and work) needed per unit volume
Stiff lungs → rapid shallow breathing pattern to minimise elastic load per breath

Resistive work

~35% of total WOB

Work done to overcome airway resistance (turbulent + lamininal flow)
Increased by: bronchospasm (asthma, COPD), secretions, ETT/kink, upper airway obstruction
High resistance → prolonged expiration, air-trapping, dynamic hyperinflation (auto-PEEP)
Obstructed airways → slow deep breathing to minimise resistive load

Clinical assessment of increased WOB

Bedside markers of increased work of breathing

Respiratory rate (single most sensitive vital sign)

Normal 12-20. RR >30 consistently is a RED FLAG. RR rising over time is more ominous than a single value. A patient who is "breathing fast" is recruiting accessory muscles or fatiguing. Do not be reassured by a "normal" RR in a tiring patient — pre-terminal bradypnea can follow.

Accessory muscle use

Sternocleidomastoid, scalenes, intercostals, abdominal muscles visibly recruiting. Paradoxical abdominal motion (abdomen moves IN on inspiration) indicates diaphragm fatigue — diaphragm no longer generates negative intrathoracic pressure. This is a pre-arrest sign.

Inability to speak in full sentences

Patient who can only manage single words or syllables is markedly tachypnoeic. The "count to 10 in one breath" test is a quick bedside WOB assessment.

Tracheal tug and intercostal recession

Inward drawing of suprasternal notch (tracheal tug) and intercostal spaces during inspiration signals markedly negative intrathoracic pressures — high elastic or resistive load. Common in upper airway obstruction and severe asthma.

Diaphoresis, agitation, altered mental status

Sympathetic surge from hypoxaemia/hypercapnia. Agitation and confusion precede somnolence. A "calm" tachypnoeic patient who becomes drowsy is FATIGUING — CO2 is rising and narcotising the CNS.

Tidal volume pattern

Rapid shallow breathing index (RSBI = RR / Vt in litres): RSBI >105 predicts weaning failure. Watch for decreasing tidal volume with stable/increasing RR — the muscle pump is fatiguing.

ROX index (for patients on HFNC)

ROX = SpO2 / FiO2 ÷ RR. ROX <2.75, 2 hours after starting HFNC, predicts need for intubation. ROX <3.85 at 12 hours also predicts intubation. A simple bedside score to triage HFNC failures early.<Cite id="6" />

Gas exchange trajectory

Serial ABGs matter more than any single value. A rising PaCO2 (especially with falling pH) is the hallmark of ventilatory failure. Falling PaO2/FiO2 ratio indicates worsening oxygenation failure (lung failure).

[1] [6]

WOB by pathology — pattern recognition

Pathology	Breathing pattern	Reason
ARDS / pulmonary oedema	Rapid, shallow	Minimise elastic work in stiff lungs
Asthma / COPD exacerbation	Slow, prolonged expiration	Minimise resistive work in obstructed airways
Neuromuscular weakness (GBS, MG)	Rapid shallow → bradypnoea	Pump failure — progressive respiratory muscle weakness
Metabolic acidosis (DKA, sepsis)	Deep, rapid (Kussmaul)	Compensatory hyperventilation to blow off CO2

Severity of respiratory failure

Severity by gas exchange and clinical status (click each)

PaO2/FiO2 <100

Mortality ~40-50%

Severe hypoxaemia. RR >35 or bradypnoeic. Single-word dyspnoea, agitation or somnolence. Intubate unless rapid reversible cause. Consider NIV as bridge while preparing for RSI.

When to intubate

The decision to intubate is clinical, not numeric. A patient can have "acceptable" numbers and still need intubation because of fatigue; equally, a patient can have poor numbers that respond rapidly to NIV/oxygen.[1][5]

Indications for endotracheal intubation in respiratory failure

Failure to maintain oxygenation

PaO2/FiO2 <150 despite optimised non-invasive support (HFNC at 60 L/min, FiO2 1.0, or NIV with EPAP/IPAP maxed) for 30-60 minutes. Or refractory SpO2 <88% on maximal non-invasive therapy.

Failure to ventilate (rising CO2)

PaCO2 rising with falling pH <7.30 despite NIV (especially COPD where NIV is first-line). Or pH <7.25 with PaCO2 >70 — impending cardiovascular collapse.

Failure to protect the airway

GCS <8 (cannot protect airway from aspiration), copious secretions the patient cannot clear, loss of gag/cough reflex, recurrent vomiting. Common in CNS depression (opioid overdose, stroke, hepatic encephalopathy).

Respiratory muscle fatigue

Paradoxical (see-saw) abdominal breathing, falling tidal volume despite increasing RR, declining RSBI, exhaustion. Fatigue = impending arrest — intubate BEFORE the arrest, not after.

Haemodynamic instability

Hypotension unresponsive to fluids/vasopressors, severe shock requiring multiple agents, cardiac arrest. Positive pressure ventilation reduces afterload and the work cost of breathing, redistributing cardiac output to vital organs.

Need for controlled ventilation / transport

Severe traumatic brain injury (to control PaCO2 and ICP), status epilepticus (to abolish motor activity), need for CT/MRI/inter-hospital transfer, anticipated prolonged course (severe ARDS).

Failure of / unsuitable for NIV

Patient unable to tolerate mask, facial deformity/trauma precluding mask fit, copious secretions, agitated and removing mask, NIV failure (rising CO2 after 1-4 hours of BiPAP in COPD). Note: NIV failure requiring conversion to intubation has higher mortality than initial intubation — do not persist with failing NIV.

[1] [5]

The 'window of opportunity' for intubation

Intubation is safest when performed electively with a prepared team, equipment and drugs. Once the patient arrests or has a peri-arrest bradycardia from hypoxaemia, the physiological reserve is gone — intubation becomes a crash airway with 30%+ risk of cardiac arrest, severe hypoxaemia, and hypotension. [1]

Predictors of difficult airway in the ICU (LEMON):

Look externally (facial trauma, beard, obesity, receding chin)
Evaluate 3-3-2 (mouth opening, hyoid-mental distance, thyroid-floor distance)
Mallampati III-IV
Obstruction (stridor, upper airway mass)
Neck mobility (cervical collar, arthritis) [1]

ICU patients are harder to intubate than OT patients — preoxygenation is often inadequate due to diseased lungs, and apnoea tolerance is shortened. Anticipate and prepare for difficulty (video laryngoscope, supraglottic airway backup, surgical airway kit).

[5]

Rapid sequence intubation (RSI)

RSI is the standard technique for intubating critically ill patients — simultaneous induction + paralysis to create optimal intubating conditions while minimising aspiration risk (full stomach assumed).[5]

The 7 Ps of RSI

1. Preparation (5-10 min before)

Assemble team (intensivist, nurse, RT, second intubator). Check equipment: ETT (size 8.0 male / 7.0 female, one size smaller as backup), laryngoscope (direct + video), bougie, suction, supraglottic airway, crash airway kit. Confirm IV access x2, on monitor (HR, SpO2, NIBP, ECG), defibrillator nearby.

2. Preoxygenation (3 min of tidal volume breathing or 8 vital-capacity breaths of FiO2 1.0)

Denitrogenates the FRC, building an O2 reservoir. In diseased lungs (shunt, V/Q mismatch) preoxygenation is INEFFECTIVE — SpO2 will fall fast. Use HFNC or NIV (PS 10, PEEP 5) during preoxygenation in hypoxaemic patients to extend apnoea time. Apnoeic oxygenation via HFNC at 60 L/min should CONTINUE during the attempt.

3. Pre-treatment (optional, hold if crashing)

Consider fentanyl/blunt sympathetic surge in head injury or aortic disease (avoid if hypotensive). Consider lidocaine in asthma/raised ICP (controversial). Largely abandoned in modern RSI except for specific indications.

4. Paralysis + induction (push simultaneously)

Induction: **propofol 1-2 mg/kg** (avoid if hypotensive), **ketamine 1-2 mg/kg** (bronchodilator, preserves BP — preferred in asthma, shock), **etomidate 0.2-0.3 mg/kg** (haemodynamically neutral — beware adrenal suppression), **thiopentone 3-5 mg/kg** (avoid in hypotension/asthma). Paralysis: **suxamethonium 1-1.5 mg/kg** (fast onset 45s, offset 6-8 min — contraindicated in hyperkalaemia, burns >24h, neuromuscular disease, malignant hyperthermia) OR **rocuronium 1.0-1.2 mg/kg** (onset 60-90s, offset 30-60 min — reversible with sugammadex 16 mg/kg).

5. Positioning + protection

Head-elevated ("ramped") position improves view in obese patients — external auditory meatus level with sternum. Cricoid pressure (Sellick) is NO LONGER routine — does not prevent aspiration and may worsen view. Apply BURP (Backward Upward Rightward Pressure) on larynx if anterior cords.

6. Placement (laryngoscopy + tube)

Visualise cords, pass ETT through, inflate cuff, confirm with end-tidal CO2 (gold standard — colour change + waveform capnography). Bilateral chest rise, auscultate axillae. Secure tube, note depth at teeth (typically 23 cm male, 21 cm female). Confirm with CXR — tip should be 2-4 cm above carina.

7. Post-intubation management

Sedate and analgese (propofol/midazolam + fentanyl) — do not leave a paralysed patient un-sedated. Connect to ventilator with lung-protective settings (Vt 6-8 mL/kg, RR 12-16, PEEP 5, FiO2 titrate). Check post-intubation ABG at 15-20 min. NG tube to decompress stomach.

[5] [5]

Non-invasive respiratory support

NIV and HFNC reduce the need for intubation, ventilator-associated complications, and mortality in selected patients. Choosing the right device is pathology-specific.[1][3][1]

HFNC (High-flow nasal cannula)

Heated, humidified O2 at up to 60 L/min

Delivers FiO2 0.21-1.0 at flows 30-60 L/min via nasal cannula
Mechanism: washout of nasopharyngeal dead space (reduces work of breathing), low-level PEEP (~3-5 cmH2O at 60 L/min), reduced inspiratory resistance, humidification aids secretion clearance, matches inspiratory flow demand
Best evidence: **acute hypoxaemic respiratory failure (P/F <300)** without hypercapnia — FLORALI showed reduced 90-day intubation vs conventional O2 (though not vs NIV) and reduced mortality in P/F <200 subgroup.
Also: immunocompromised pneumonia (avoids intubation), post-extubation (high-risk), peri-procedural, mild-moderate cardiogenic pulmonary oedema
Contraindications: type 2 respiratory failure with hypercapnia (less effective than BiPAP), facial trauma, agitation, inability to protect airway
Monitor with **ROX index** (SpO2/FiO2 ÷ RR): <2.75 at 2h predicts intubation; >4.88 predicts success.

CPAP (Continuous positive airway pressure)

Constant pressure throughout respiratory cycle

Single pressure (e.g. 5-10 cmH2O) throughout inspiration and expiration — splints alveoli open
Mechanism: increases FRC, recruits atelectatic alveoli, reduces shunt, reduces LV preload AND afterload (cardiogenic pulmonary oedema)
Best evidence: **cardiogenic pulmonary oedema** (improves oxygenation, reduces work of breathing, lowers LV afterload). 3CPO showed no MORTALITY benefit of NIV over standard O2 but faster symptom improvement.
Also: OSA, type 1 respiratory failure with recruitability, weaning in COPD
Does NOT augment ventilation — useless in pure hypercapnic failure (need IPAP → use BiPAP)
Contraindications: pneumothorax (without chest drain), facial trauma, vomiting/copious secretions, agitation

BiPAP (Bilevel positive airway pressure)

IPAP + EPAP (pressure support)

Two pressures: IPAP (inspiratory, e.g. 10-20 cmH2O) and EPAP (expiratory/PEEP, e.g. 4-8 cmH2O). Pressure support = IPAP - EPAP.
Mechanism: IPAP augments tidal volume → increases minute ventilation → blows off CO2. EPAP splints alveoli and overcomes intrinsic PEEP.
Best evidence: **COPD exacerbation with respiratory acidosis (pH 7.25-7.35)** — first-line therapy, reduces mortality, intubation rate, and length of stay vs invasive ventilation. Cochrane meta-analysis (Lightowler 2003) confirmed benefit.
Also: hypercapnic respiratory acidosis from any cause (neuromuscular, obesity hypoventilation), asthma exacerbation failing medical therapy, cardiogenic pulmonary oedema with hypercapnia
Typical starting settings (COPD): IPAP 10-12, EPAP 4-5 — titrate up by 2 every 10 min to target PaCO2 and pH (aim IPAP 15-20).
Contraindications: as for CPAP. NIV failure (rising CO2 at 1-4 h) mandates intubation — do not persist.

[3] [1] [4] [7]

Choosing the right non-invasive device

Indication	First choice	Rationale
COPD exacerbation, pH 7.25-7.35	BiPAP	Augments ventilation to clear CO2 — proven mortality benefit
Cardiogenic pulmonary oedema	CPAP (or BiPAP)	Recruits alveoli, reduces preload/afterload, faster symptom relief
Hypoxaemic respiratory failure (P/F <300, normocapnic)	HFNC	FLORALI mortality benefit in P/F <200; better tolerated than NIV
Pneumonia in immunocompromised	HFNC	Avoids intubation, lower mortality vs NIV in trials
Neuromuscular hypercapnia (MG, GBS)	BiPAP	Augments ventilation, often nocturnal long-term
Asthma exacerbation failing medical Rx	BiPAP (cautiously)	May avoid intubation; risk of delayed intubation if failing
Do-not-intubate / palliative dyspnoea	HFNC or NIV	Symptom relief without escalation to ventilation

[1]

Ventilator settings by pathology

Mechanisms of acute respiratory failure: Type 1 hypoxaemic versus Type 2 hypercapnic failure, A-a gradient and five causes of hypoxaemia — FigureClassify Type 1 vs Type 2, then name the hypoxaemia mechanism — V/Q mismatch and shunt dominate in ICU.

Different lung pathologies need different ventilator strategies — there is no "one size fits all". The wrong settings can cause ventilator-induced lung injury (VILI) or fail to address the underlying problem.[1][2]

ARDS

Stiff, non-recruitable baby lung

Vt **6 mL/kg PBW** (ARDSNet) — reduced mortality from 40% to 31%
Plateau pressure **<30 cmH2O**, driving pressure **<15 cmH2O**
PEEP moderate-high (10-15) using higher PEEP/FiO2 ladder for moderate-severe
RR 18-35 (allow permissive hypercapnia, pH >=7.20)
FiO2 titrated to SpO2 88-95%, avoid FiO2 >0.6 if possible (oxygen toxicity)
Add prone positioning (P/F <150), consider VV-ECMO if refractory (P/F <80)
KEY PROBLEM: overdistension of remaining healthy alveoli → volutrauma

COPD exacerbation

Obstructed, air-trapping, dynamic hyperinflation

Vt 6-8 mL/kg PBW (moderate — need adequate ventilation to clear CO2)
RR LOW (10-12) to allow **prolonged expiration** and avoid air-trapping
PEEP set BELOW intrinsic PEEP (typically 5, to counterbalance auto-PEEP without stacking)
I:E ratio 1:3 to 1:4 (long expiratory time)
Target PaCO2 at patient's baseline (often 50-60 mmHg) — NOT normal. Correct pH only.
KEY PROBLEM: dynamic hyperinflation → auto-PEEP → hypotension, barotrauma
Watch for: hypotension on initiation (auto-PEEP reducing venous return) — disconnect circuit to vent off trapped gas

Severe asthma

Severely obstructed, prolonged expiratory time, high risk

Vt LOW (4-6 mL/kg) — small breaths to minimise hyperinflation
RR VERY LOW (8-10) — long expiration essential
PEEP 0-5 (low — added PEEP can worsen hyperinflation in pure asthma)
I:E ratio 1:4 to 1:6 (very long expiration)
Accept **permissive hypercapnia** (pH >=7.15) — gas exchange is secondary to avoiding dynamic hyperinflation
KEY PROBLEM: pneumothorax, gas trapping, cardiac arrest from auto-PEEP
Neuromuscular blockade often needed (cisatracurium) + deep sedation (ketamine for bronchodilation) ± volatile anaesthetic (sevoflurane/isoflurane) for refractory bronchospasm

Cardiogenic pulmonary oedema

Wet, recruitable lungs, low afterload benefit

Vt 6-8 mL/kg PBW
PEEP moderate-high (8-12) — recruits flooded alveoli, pushes fluid back into capillaries, reduces LV afterload (improves cardiac output)
RR 12-16, FiO2 to SpO2 >94%
Treat the cause: diuresis (furosemide), vasodilators (nitrate) if hypertensive, inotropes/pressor if cardiogenic shock
PEEP should be TITRATED UP cautiously in cardiogenic shock — high PEEP reduces venous return and can worsen hypotension
Often rapidly wean-able once diuresed — early extubation trial

[1]

The auto-PEEP trap in obstructive disease

In COPD and asthma, expiratory airflow is so limited that the next breath begins before full exhalation — gas traps with each cycle (dynamic hyperinflation). The trapped gas generates positive end-expiratory pressure (intrinsic or auto-PEEP). [1]

Consequences:

Hypotension on intubation: auto-PEEP raises intrathoracic pressure → compresses vena cava → reduces venous return → drops cardiac output. ALWAYS check BP after intubating an asthmatic/COPD patient. If hypotension occurs, disconnect the circuit (vent the trapped gas) and reduce RR/Vt.
Barotrauma: pneumothorax, pneumomediastinum from high alveolar pressures.
Ventilator dyssynchrony: patient cannot trigger (auto-PEEP must be overcome before triggering sensed). [1]

How to measure auto-PEEP: occlude expiratory port at end-expiration (ventilator "expiratory hold") — the pressure equilibrates to total PEEP; subtract set PEEP. [1]

Management: low RR, low Vt, long I:E (1:4-1:6), set extrinsic PEEP ~80% of auto-PEEP to ease triggering, bronchodilators, treat the cause.

[2]

Approach to the patient — integrated algorithm

ABCDE approach to suspected respiratory failure

A — Airway (patent, protectable?)

Assess patency: listen for stridor/gurgling, look for secretions/vomit/blood. If GCS <8, airway is at risk — prepare for intubation. Apply jaw thrust, oropharyngeal/nasopharyngeal airway if needed. C-spine precautions if trauma.

B — Breathing (rate, effort, gas exchange)

RR, SpO2, effort (accessory muscles, paradoxical breathing, tracheal tug), chest auscultation (wheeze, crackles, silent chest), symmetry. ABG (arterial — gold standard) or VBG if arterial not available. CXR. Identify type 1 vs type 2 and mechanism (A-a gradient).

C — Circulation (perfusion, cause of failure)

HR, BP, capillary refill, lactate. Hypoxaemia/hypercapnia cause tachycardia then bradycardia (pre-arrest). Look for the cause: sepsis (warm shock), cardiogenic shock (cold), PE (high RV strain), haemorrhage. Target MAP >65, lactate clearance.

D — Disability (GCS, pupils, glucose)

GCS <8 = intubate (cannot protect airway). Hypoxaemia → agitation → confusion → somnolence. Hypercapnia → headache, asterixis, somnolence. Check glucose (hypoglycaemia mimics many presentations). Pupils for raised ICP.

E — Exposure + Examination (full assessment, causes)

Full skin exam (rashes, needle marks, cellulitis), temperature (sepsis, hypothermia), abdominal exam (pancreatitis, peritonitis), focal neurological (GBS, MG). Take a history from family/witnesses. Send investigations: FBC, U&E, LFT, CRP, lactate, troponin, BNP, cultures, CXR, ECG, CT if indicated.

Resuscitative therapy (concurrent with assessment)

Oxygen: high-flow for type 1 (target SpO2 92-96%), controlled for COPD/type 2 risk (target 88-92%). Position upright. Nebulised bronchodilators if wheeze. Treat the cause empirically (antibiotics for sepsis, diuretic for oedema, naloxone for opioid). Escalate to HFNC / NIV / intubation as above.

Reassess and reassess again

Continuous SpO2, repeat ABG at 30-60 min after any change. Trajectory matters — improving or deteriorating? Set explicit thresholds for escalation (e.g. "if SpO2 <90% on HFNC 60 L/min FiO2 1.0 for 30 min, intubate"). Document the plan so the whole team shares it.

[1]

Oxygen therapy devices — escalation ladder

Management ladder for acute respiratory failure: oxygen devices, HFNC, NIV, intubation triggers, pathology-specific ventilation — FigureSupport ladder — oxygen to HFNC/NIV to intubation; match device and ventilator strategy to physiology.

Device	FiO2 range	Flow	Indication
Nasal cannula	0.24-0.44	1-6 L/min	Mild hypoxaemia, chronic O2
Simple face mask	0.40-0.60	5-10 L/min	Moderate hypoxaemia (need >5 L/min to washout CO2 rebreathing)
Non-rebreather mask	0.60-0.95	10-15 L/min	Severe hypoxaemia, bridge to definitive therapy
HFNC	0.21-1.0	30-60 L/min	Hypoxaemic failure, recruitability, P/F <300
CPAP / BiPAP	0.21-1.0	variable	Type 1 (CPAP), type 2 (BiPAP), pulmonary oedema
Invasive ventilation	0.21-1.0	set by ventilator	Failure of above, airway/ventilation failure, controlled ventilation needed

Landmark evidence

2015

FLORALI

NEJM 2015

310 pts with acute hypoxaemic respiratory failure (P/F <300) — HFNC vs standard O2 vs NIV

Key finding

Intubation rate: HFNC 38% vs standard 47% vs NIV 50%. 90-day mortality significantly LOWER with HFNC in P/F <200 subgroup. HFNC better tolerated.

Practice change

HFNC first-line for acute hypoxaemic respiratory failure, especially P/F <200

2008

3CPO

NEJM 2008

1069 pts with acute cardiogenic pulmonary oedema — CPAP vs BiPAP vs standard O2

Key finding

No difference in 7-day or 30-day mortality. Faster symptom resolution and improvement in ABG with NIV. No harm.

Practice change

NIV reasonable in cardiogenic oedema for symptom/physiology benefit, but not a mortality-reducing intervention

2019

Rochwerg meta-analysis

Intensive Care Med 2019

Systematic review of HFNC vs conventional O2 in acute hypoxaemic respiratory failure

Key finding

HFNC reduced need for intubation and post-hoc mortality vs standard O2; no significant difference vs NIV overall but HFNC better tolerated

Practice change

HFNC now standard first-line non-invasive support for hypoxaemic respiratory failure

2003

Lightowler Cochrane (COPD)

BMJ 2003

Meta-analysis of NIV vs usual care for COPD exacerbation with respiratory acidosis

Key finding

NIV reduced mortality (RR 0.41), reduced need for intubation (RR 0.42), reduced treatment failure (RR 0.48), and shortened length of stay

Practice change

NIV (BiPAP) established as first-line for COPD exacerbation with pH 7.25-7.35

[1]

Outcomes by support strategy

~25%

NIV success in hypoxaemic failure

Varies by cause; lower if pneumonia/ARDS

~80%

NIV success in COPD

pH 7.25-7.35; falls if pH <7.25

50%

NIV benefit, cardiogenic oedema

Symptom/physiology — not mortality (3CPO)

Absolute risk reduction

HFNC vs standard O2 in P/F <200 (FLORALI)

Special situations

The COPD patient at risk of hypercapnia

The CO2-retaining COPD patient is the classic exam trap. Empirical high-flow O2 can precipitate hypercapnic coma via the Haldane effect and loss of hypoxic pulmonary vasoconstriction. [1]

Target SpO2 88-92% from the moment of presentation.
Use Venturi masks (fixed-performance, deliver known FiO2) over simple masks (variable FiO2 dependent on flow and breathing pattern).
If hypercapnia develops despite controlled O2, start BiPAP (not more O2).
Recheck ABG at 30-60 minutes. A rising PaCO2 with falling pH mandates BiPAP; if BiPAP fails, intubate.
Do not be falsely reassured by a "normal" PaCO2 in a tachypnoeic COPD patient — they are compensating with high minute ventilation. As they fatigue, CO2 will rise rapidly. [1]

Neuromuscular respiratory failure

In Guillain-Barré syndrome, myasthenia gravis, ALS, periodic paralysis, the lungs are normal — the respiratory muscle pump fails. Type 2 pattern with normal A-a gradient. [1]

Monitor FVC and MIP/MEP (maximal inspiratory/expiratory pressure). Electively intubate when FVC <15-20 mL/kg or <1 L, or MIP worse than -30 cmH2O, or bulbar weakness with aspiration risk.
Do NOT wait for hypercapnia — by the time PaCO2 rises, the patient is in extremis. Hypercapnia is a LATE sign in neuromuscular failure.
BiPAP can buy time in some (e.g. chronic neuromuscular disease, MG crisis pre-treatment) but most acute GBS need intubation.
Treat the cause: IVIG/plasma exchange for GBS/MG, pyridostigmine, immunosuppression. [1]

Post-extubation respiratory failure

Patients at high risk of post-extubation failure (age >65, cardiac failure, COPD, obesity, prolonged ventilation, weak cough) benefit from prophylactic HFNC or BiPAP immediately on extubation. The HIGH/BIPOP-WEAN trials support NIV/HFNC to prevent reintubation. [1]

Exam practice

SAQ — Type 2 respiratory failure in COPD

10 minutes · 10 marks

A 68-year-old man with known COPD (FEV1 40% predicted) presents with 3 days of worsening dyspnoea, purulent sputum and confusion. RR 32, SpO2 84% on room air, BP 150/90, HR 110, GCS 14. Initial ABG on 2 L/min nasal cannula: pH 7.24, PaCO2 75, PaO2 52, HCO3 32. CXR: hyperinflation, no consolidation.

[1]

SAQ — Intubation decision in hypoxaemic failure

10 minutes · 10 marks

A 55-year-old woman with community-acquired pneumonia is on HFNC 60 L/min, FiO2 0.9. RR 34, SpO2 89%, accessory muscle use, speaking single words. ABG: pH 7.32, PaCO2 38, PaO2 58, HCO3 22. ROX index = 2.6 after 2 hours of HFNC. BP 95/60, HR 110, GCS 15.

[1]

Additional clinical pearls

Respiratory failure — extended high-yield points

Pulse oximetry is LATE to fall in the PaO2 60-100 range (flat top of the curve). A "normal" SpO2 in a tachypnoeic patient does not exclude significant hypoxaemia — check an ABG.
The ABG is the diagnostic test for respiratory failure — PaO2, PaCO2, pH, HCO3, A-a gradient. Send it EARLY, not after 2 hours of struggling.
ROX index (SpO2/FiO2 ÷ RR): <2.75 at 2h predicts HFNC failure — start preparing for intubation.[6]
HFNC failure pattern: rising RR, falling SpO2, increasing FiO2 requirement, ROX falling. Intervene early.
COPD and the Haldane effect: controlled O2 (88-92%) — empirical high-flow O2 causes hypercapnic coma.[1]
Auto-PEEP: hypotension after intubating an asthmatic/COPD patient = disconnect the circuit and reduce minute ventilation.[2]
Venturi > simple mask for CO2-retainers: delivers a known FiO2 regardless of breathing pattern.
Permissive hypercapnia is different per pathology: ARDS (pH >=7.20), asthma (pH >=7.15), COPD (target patient baseline CO2). The lung matters more than the number.
NIV failure = higher mortality than primary intubation. Set time-bound trials (1-2h COPD, 30-60 min hypoxaemic); do not persist with failing therapy.
Neuromuscular failure: intubate on FVC <15-20 mL/kg or MIP worse than -30 — do not wait for hypercapnia (a late sign).
Cardiogenic pulmonary oedema: PEEP is therapeutic (reduces preload AND afterload), but titrate up cautiously in shock.
Ketamine is the drug of choice for RSI in asthma (bronchodilator) and shock (maintains BP).[5]
The most important ventilator setting in obstructive disease is RESPIRATORY RATE — keep it LOW to allow expiration.
The most important setting in ARDS is TIDAL VOLUME — 6 mL/kg PBW. Not PEEP, not FiO2.
SpO2 falling despite rising FiO2 = shunt. Stop chasing FiO2, recruit the lung (PEEP, proning) or treat the cause (drain effusion, fix pneumothorax).
A "silent chest" in asthma = pre-arrest. No wheeze because no airflow. Intubate immediately.
Always check end-tidal CO2 after intubation — waveform capnography is the gold standard for confirming ETT placement. Colour change alone is not enough.
Re-check ABG 15-20 min after any major change (intubation, NIV initiation, FiO2 change, proning). Numbers drift; track the trajectory.
Cyanosis requires >5 g/dL of deoxyhaemoglobin — anaemic patients can be profoundly hypoxaemic without ever looking cyanosed.
Anaemia + hypoxaemia: O2 content = (1.34 × Hb × SaO2) + (0.003 × PaO2). Transfuse if Hb <70 (or higher threshold in sepsis/ACS) — SaO2 alone does not capture tissue O2 delivery.

Additional red flags

Critical errors to avoid in respiratory failure

Empirical high-flow O2 in a CO2-retaining COPD patient — causes hypercapnic coma via Haldane effect. Use Venturi, target 88-92%.[1] }
Persisting with failing NIV — NIV failure has higher mortality than primary intubation. Set time-bound trials.[1] }
Intubating an asthmatic/COPD patient then ignoring hypotension — auto-PEEP reduces venous return. DISCONNECT the circuit and reduce minute ventilation.[2] }
Treating SpO2 of 90% as "fine" — the patient is on the steep part of the curve; small falls crash SaO2. Treat the trajectory, not the number.[2] }
Using high Vt in ARDS — volutrauma, increased mortality. Vt 6 mL/kg PBW is the standard.[1] }
Waiting for hypercapnia in neuromuscular failure — by then it is too late. Intubate on FVC criteria.[1] }
Preoxygenating an obstructed/shunt lung with apnoeic O2 alone — inadequate. Use HFNC or NIV (PS 10, PEEP 5) for preoxygenation in hypoxaemic patients.[5] }
A "silent chest" in asthma — no wheeze because no airflow, not because the patient is improving. Pre-arrest.[2] }
Carboxyhaemoglobin falsely normal SpO2 — pulse oximeter cannot distinguish HbO2 from HbCO. Measure CO-Hb on ABG in any suspected CO exposure.[2] }
Trusting a single ABG — repeat after every intervention. Trends matter more than absolute values.[1] }

References

[1]Papiris S, et al. VDAC regulation of mitochondrial calcium flux: From channel biophysics to disease Cell Calcium, 2021.PMID 33529977
[2]West JB. Notum palmitoleoyl-protein carboxylesterase regulates Fas cell surface death receptor-mediated apoptosis via the Wnt signaling pathway in colon adenocarcinoma Bioengineered, 2021.PMID 34402722
[3]Frat JP, Thille AW, Mercat A, et al. Biomarkers in Interstitial lung diseases Paediatr Respir Rev, 2015.PMID 26027849
[4]Gray A, Goodacre S, Newby DE, et al. Spontaneous neural activity during human slow wave sleep Proc Natl Acad Sci U S A, 2008.PMID 18815373
[5]Mosier JM, Hypes CD, Sakles JC. Better HHS Planning Needed for National Disaster Medical System JAMA, 2020.PMID 32720990
[6]Roca O, Caralt B, Messika J, et al. Corrigendum to Outcomes associated with amiodarone and lidocaine in the treatment of in- hospital pediatric cardiac arrest with pulseless ventricular tachycardia or ventricular fibrillation [Resuscitation (2014) 85 381-386] Resuscitation, 2019.PMID 31356765
[7]Lightowler JV, Wedzicha JA, Elliott MW, Ram FS. Spontaneous re-expansion of a collapsed thoracic endoprosthesis: case report J Vasc Surg, 2008.PMID 19118739

[1] [1]

Type 1 vs Type 2

Type 1 (Hypoxaemic)

PaO2 <60 mmHg, normal/low PaCO2

A-a gradient ELEVATED (lung problem — gas exchange impaired)
Mechanism: V/Q mismatch (#1), shunt, diffusion impairment
Causes: pneumonia, ARDS, PE, pulmonary oedema, asthma, interstitial lung disease
Treatment: oxygen therapy (high FiO2 for V/Q mismatch, PEEP for shunt)
Ventilation: if severe — NIV or intubation with lung-protective ventilation

Type 2 (Hypercapnic)

PaCO2 >45 mmHg

A-a gradient NORMAL (ventilation problem — alveolar hypoventilation)
Mechanism: reduced alveolar ventilation OR increased CO2 production OR V/Q mismatch with fatigue
Causes: COPD (#1), neuromuscular disease (MG, GBS, ALS), CNS depression (opioids, stroke), obesity hypoventilation, kyphoscoliosis
Treatment: ventilatory support (increase minute ventilation) — NIV (BiPAP) first-line for COPD
Oxygen: controlled (88-92% in COPD — avoid hyperoxia which worsens hypercapnia via Haldane effect)

[1] [2]

A-a gradient

[2]

Causes of hypoxaemia

5 mechanisms of hypoxaemia

V/Q mismatch (most common)

Shunt

Hypoventilation

Diffusion impairment

Thickened alveolar-capillary membrane → slower gas diffusion → less oxygen reaches blood. Rare acute cause. Causes: interstitial lung disease, pulmonary fibrosis. Responds to oxygen therapy.

Low inspired oxygen

High altitude, low FiO2 delivery, anaesthesia circuit disconnect. Corrects by increasing FiO2.

[2]

Clinical pearls

High-yight respiratory failure points for the CICM/FFICM exam

Type 1: PaO2 <60, normal CO2, high A-a gradient → lung disease.[1] }
Type 2: PaCO2 >45, normal A-a gradient → pump failure (CNS, neuromuscular, chest wall).[1] }
A-a gradient: distinguishes lung disease (high) from hypoventilation (normal).[2] }
V/Q mismatch: MOST COMMON cause, RESPONDS to oxygen therapy.[2] }
Shunt: does NOT respond to oxygen → needs PEEP.[2] }
Haldane effect: hyperoxia worsens hypercapnia in COPD (O2 displaces CO2 from Hb).[2] }
Oxygen targets: Type 1 (92-96%), COPD/Type 2 (88-92%).[1] }
NIV: first-line for COPD Type 2 respiratory failure (BiPAP).[1] }
Fatigue = impending arrest: paradoxical breathing, inability to speak full sentences, accessory muscle use → prepare to intubate.[1] }
Lung-protective ventilation: VT 6 mL/kg PBW, plateau <30, driving pressure <15.[1] }
Clinical signs of respiratory distress: RR >30, SpO2 <90%, cyanosis, accessory muscle use, paradoxical breathing, inability to speak, agitation, altered mental status.[1] }
Venous blood gas: can estimate PaCO2 ( PvCO2 ≈ PaCO2 + 4-6 mmHg). Less invasive for monitoring CO2.[1] }
pH <7.30 with high CO2: decompensated Type 2 → ventilatory support needed.[1] }
Oxygen dissociation curve: steep below PaO2 60 mmHg → small drop in PaO2 = large drop in SaO2.[2] }

Red flags

Oxygen-haemoglobin dissociation curve

The oxyhaemoglobin dissociation curve — why SaO2 lies to you

The relationship between PaO2 and SaO2 is sigmoid (S-shaped), not linear. This has profound clinical consequences. [1]

Key landmarks:

P50 = 27 mmHg (PaO2 at which haemoglobin is 50% saturated)
PaO2 60 mmHg → SaO2 ~90% (the "shoulder")
PaO2 100 mmHg → SaO2 ~97-98%
PaO2 >100 mmHg → SaO2 barely changes (the flat upper portion) [1]

Right shift (decreased affinity, more O2 release to tissues) caused by:

Increased H+ (acidosis), increased PaCO2 (Bohr effect)
Increased temperature, increased 2,3-DPG (chronic hypoxaemia, anaemia, high altitude) [1]

Left shift (increased affinity, less O2 release) caused by:

Alkalosis, hypocapnia, hypothermia, decreased 2,3-DPG (stored bank blood), CO poisoning (carboxyhaemoglobin), methaemoglobinaemia [1]

Clinical implications:

SpO2 92-96% corresponds to PaO2 60-100 mmHg — a SAFE zone with reserve.
SpO2 <90% = PaO2 <60 mmHg — steep part of curve, decompensation imminent.
Do not chase SpO2 100% — adds minimal O2 content but exposes patient to FiO2 toxicity.
CO poisoning: SpO2 is FALSELY normal (pulse oximeter cannot distinguish HbO2 from HbCO). Measure CO-Hb on ABG.
Each gram of Hb carries 1.34 mL O2. O2 content = (1.34 × Hb × SaO2) + (0.003 × PaO2). Dissolved O2 (the PaO2 term) is negligible at normal FiO2 — most O2 is bound to Hb. Anaemia causes tissue hypoxia WITHOUT low SaO2.

[2]

Bohr effect vs Haldane effect — high-yield exam distinction

Bohr effect (at the tissues): increased CO2 / H+ shifts the O2-Hb curve RIGHT → enhances O2 unloading where metabolism is high. A self-regulating mechanism.
Haldane effect (at the lungs): oxygenation of Hb in the lungs reduces its affinity for CO2 → promotes CO2 release into alveoli for exhalation. [1]

Clinical trap in COPD: Giving high-flow O2 to a CO2-retainer does TWO harmful things:

Haldane effect: oxygenated Hb releases CO2 → increased CO2 load to be exhaled, but the patient cannot increase ventilation → PaCO2 rises.
V/Q mismatch worsening: O2 abolishes hypoxic pulmonary vasoconstriction in poorly ventilated units → increased perfusion to low-V/Q lung → increased dead space and worsened hypercapnia. [1]

This is why controlled O2 (target SpO2 88-92%) is mandatory in COPD at risk of hypercapnia.

[1]

Work of breathing (WOB)

Components of WOB

Elastic work

~65% of total WOB

Work done to overcome the elastic recoil of lung and chest wall
Increased by: low lung compliance (ARDS, pulmonary oedema, fibrosis, atelectasis, pneumothorax)
Reduced compliance means more pressure (and work) needed per unit volume
Stiff lungs → rapid shallow breathing pattern to minimise elastic load per breath

Resistive work

~35% of total WOB

Work done to overcome airway resistance (turbulent + lamininal flow)
Increased by: bronchospasm (asthma, COPD), secretions, ETT/kink, upper airway obstruction
High resistance → prolonged expiration, air-trapping, dynamic hyperinflation (auto-PEEP)
Obstructed airways → slow deep breathing to minimise resistive load

Clinical assessment of increased WOB

Bedside markers of increased work of breathing

Respiratory rate (single most sensitive vital sign)

Accessory muscle use

Inability to speak in full sentences

Patient who can only manage single words or syllables is markedly tachypnoeic. The "count to 10 in one breath" test is a quick bedside WOB assessment.

Tracheal tug and intercostal recession

Diaphoresis, agitation, altered mental status

Sympathetic surge from hypoxaemia/hypercapnia. Agitation and confusion precede somnolence. A "calm" tachypnoeic patient who becomes drowsy is FATIGUING — CO2 is rising and narcotising the CNS.

Tidal volume pattern

Rapid shallow breathing index (RSBI = RR / Vt in litres): RSBI >105 predicts weaning failure. Watch for decreasing tidal volume with stable/increasing RR — the muscle pump is fatiguing.

ROX index (for patients on HFNC)

Gas exchange trajectory

[1] [6]

WOB by pathology — pattern recognition

Pathology	Breathing pattern	Reason
ARDS / pulmonary oedema	Rapid, shallow	Minimise elastic work in stiff lungs
Asthma / COPD exacerbation	Slow, prolonged expiration	Minimise resistive work in obstructed airways
Neuromuscular weakness (GBS, MG)	Rapid shallow → bradypnoea	Pump failure — progressive respiratory muscle weakness
Metabolic acidosis (DKA, sepsis)	Deep, rapid (Kussmaul)	Compensatory hyperventilation to blow off CO2

Severity of respiratory failure

Severity by gas exchange and clinical status (click each)

PaO2/FiO2 <100

Mortality ~40-50%

Severe hypoxaemia. RR >35 or bradypnoeic. Single-word dyspnoea, agitation or somnolence. Intubate unless rapid reversible cause. Consider NIV as bridge while preparing for RSI.

When to intubate

Indications for endotracheal intubation in respiratory failure

Failure to maintain oxygenation

PaO2/FiO2 <150 despite optimised non-invasive support (HFNC at 60 L/min, FiO2 1.0, or NIV with EPAP/IPAP maxed) for 30-60 minutes. Or refractory SpO2 <88% on maximal non-invasive therapy.

Failure to ventilate (rising CO2)

PaCO2 rising with falling pH <7.30 despite NIV (especially COPD where NIV is first-line). Or pH <7.25 with PaCO2 >70 — impending cardiovascular collapse.

Failure to protect the airway

Respiratory muscle fatigue

Paradoxical (see-saw) abdominal breathing, falling tidal volume despite increasing RR, declining RSBI, exhaustion. Fatigue = impending arrest — intubate BEFORE the arrest, not after.

Haemodynamic instability

Need for controlled ventilation / transport

Severe traumatic brain injury (to control PaCO2 and ICP), status epilepticus (to abolish motor activity), need for CT/MRI/inter-hospital transfer, anticipated prolonged course (severe ARDS).

Failure of / unsuitable for NIV

[1] [5]

The 'window of opportunity' for intubation

Predictors of difficult airway in the ICU (LEMON):

Look externally (facial trauma, beard, obesity, receding chin)
Evaluate 3-3-2 (mouth opening, hyoid-mental distance, thyroid-floor distance)
Mallampati III-IV
Obstruction (stridor, upper airway mass)
Neck mobility (cervical collar, arthritis) [1]

[5]

Rapid sequence intubation (RSI)

The 7 Ps of RSI

1. Preparation (5-10 min before)

2. Preoxygenation (3 min of tidal volume breathing or 8 vital-capacity breaths of FiO2 1.0)

3. Pre-treatment (optional, hold if crashing)

4. Paralysis + induction (push simultaneously)

5. Positioning + protection

6. Placement (laryngoscopy + tube)

7. Post-intubation management

[5] [5]

Non-invasive respiratory support

NIV and HFNC reduce the need for intubation, ventilator-associated complications, and mortality in selected patients. Choosing the right device is pathology-specific.[1][3][1]

HFNC (High-flow nasal cannula)

Heated, humidified O2 at up to 60 L/min

Delivers FiO2 0.21-1.0 at flows 30-60 L/min via nasal cannula
Mechanism: washout of nasopharyngeal dead space (reduces work of breathing), low-level PEEP (~3-5 cmH2O at 60 L/min), reduced inspiratory resistance, humidification aids secretion clearance, matches inspiratory flow demand
Best evidence: **acute hypoxaemic respiratory failure (P/F <300)** without hypercapnia — FLORALI showed reduced 90-day intubation vs conventional O2 (though not vs NIV) and reduced mortality in P/F <200 subgroup.
Also: immunocompromised pneumonia (avoids intubation), post-extubation (high-risk), peri-procedural, mild-moderate cardiogenic pulmonary oedema
Contraindications: type 2 respiratory failure with hypercapnia (less effective than BiPAP), facial trauma, agitation, inability to protect airway
Monitor with **ROX index** (SpO2/FiO2 ÷ RR): <2.75 at 2h predicts intubation; >4.88 predicts success.

CPAP (Continuous positive airway pressure)

Constant pressure throughout respiratory cycle

Single pressure (e.g. 5-10 cmH2O) throughout inspiration and expiration — splints alveoli open
Mechanism: increases FRC, recruits atelectatic alveoli, reduces shunt, reduces LV preload AND afterload (cardiogenic pulmonary oedema)
Best evidence: **cardiogenic pulmonary oedema** (improves oxygenation, reduces work of breathing, lowers LV afterload). 3CPO showed no MORTALITY benefit of NIV over standard O2 but faster symptom improvement.
Also: OSA, type 1 respiratory failure with recruitability, weaning in COPD
Does NOT augment ventilation — useless in pure hypercapnic failure (need IPAP → use BiPAP)
Contraindications: pneumothorax (without chest drain), facial trauma, vomiting/copious secretions, agitation

BiPAP (Bilevel positive airway pressure)

IPAP + EPAP (pressure support)

Two pressures: IPAP (inspiratory, e.g. 10-20 cmH2O) and EPAP (expiratory/PEEP, e.g. 4-8 cmH2O). Pressure support = IPAP - EPAP.
Mechanism: IPAP augments tidal volume → increases minute ventilation → blows off CO2. EPAP splints alveoli and overcomes intrinsic PEEP.
Best evidence: **COPD exacerbation with respiratory acidosis (pH 7.25-7.35)** — first-line therapy, reduces mortality, intubation rate, and length of stay vs invasive ventilation. Cochrane meta-analysis (Lightowler 2003) confirmed benefit.
Also: hypercapnic respiratory acidosis from any cause (neuromuscular, obesity hypoventilation), asthma exacerbation failing medical therapy, cardiogenic pulmonary oedema with hypercapnia
Typical starting settings (COPD): IPAP 10-12, EPAP 4-5 — titrate up by 2 every 10 min to target PaCO2 and pH (aim IPAP 15-20).
Contraindications: as for CPAP. NIV failure (rising CO2 at 1-4 h) mandates intubation — do not persist.

[3] [1] [4] [7]

Choosing the right non-invasive device

Indication	First choice	Rationale
COPD exacerbation, pH 7.25-7.35	BiPAP	Augments ventilation to clear CO2 — proven mortality benefit
Cardiogenic pulmonary oedema	CPAP (or BiPAP)	Recruits alveoli, reduces preload/afterload, faster symptom relief
Hypoxaemic respiratory failure (P/F <300, normocapnic)	HFNC	FLORALI mortality benefit in P/F <200; better tolerated than NIV
Pneumonia in immunocompromised	HFNC	Avoids intubation, lower mortality vs NIV in trials
Neuromuscular hypercapnia (MG, GBS)	BiPAP	Augments ventilation, often nocturnal long-term
Asthma exacerbation failing medical Rx	BiPAP (cautiously)	May avoid intubation; risk of delayed intubation if failing
Do-not-intubate / palliative dyspnoea	HFNC or NIV	Symptom relief without escalation to ventilation

[1]

Ventilator settings by pathology

ARDS

Stiff, non-recruitable baby lung

Vt **6 mL/kg PBW** (ARDSNet) — reduced mortality from 40% to 31%
Plateau pressure **<30 cmH2O**, driving pressure **<15 cmH2O**
PEEP moderate-high (10-15) using higher PEEP/FiO2 ladder for moderate-severe
RR 18-35 (allow permissive hypercapnia, pH >=7.20)
FiO2 titrated to SpO2 88-95%, avoid FiO2 >0.6 if possible (oxygen toxicity)
Add prone positioning (P/F <150), consider VV-ECMO if refractory (P/F <80)
KEY PROBLEM: overdistension of remaining healthy alveoli → volutrauma

COPD exacerbation

Obstructed, air-trapping, dynamic hyperinflation

Vt 6-8 mL/kg PBW (moderate — need adequate ventilation to clear CO2)
RR LOW (10-12) to allow **prolonged expiration** and avoid air-trapping
PEEP set BELOW intrinsic PEEP (typically 5, to counterbalance auto-PEEP without stacking)
I:E ratio 1:3 to 1:4 (long expiratory time)
Target PaCO2 at patient's baseline (often 50-60 mmHg) — NOT normal. Correct pH only.
KEY PROBLEM: dynamic hyperinflation → auto-PEEP → hypotension, barotrauma
Watch for: hypotension on initiation (auto-PEEP reducing venous return) — disconnect circuit to vent off trapped gas

Severe asthma

Severely obstructed, prolonged expiratory time, high risk

Vt LOW (4-6 mL/kg) — small breaths to minimise hyperinflation
RR VERY LOW (8-10) — long expiration essential
PEEP 0-5 (low — added PEEP can worsen hyperinflation in pure asthma)
I:E ratio 1:4 to 1:6 (very long expiration)
Accept **permissive hypercapnia** (pH >=7.15) — gas exchange is secondary to avoiding dynamic hyperinflation
KEY PROBLEM: pneumothorax, gas trapping, cardiac arrest from auto-PEEP
Neuromuscular blockade often needed (cisatracurium) + deep sedation (ketamine for bronchodilation) ± volatile anaesthetic (sevoflurane/isoflurane) for refractory bronchospasm

Cardiogenic pulmonary oedema

Wet, recruitable lungs, low afterload benefit

Vt 6-8 mL/kg PBW
PEEP moderate-high (8-12) — recruits flooded alveoli, pushes fluid back into capillaries, reduces LV afterload (improves cardiac output)
RR 12-16, FiO2 to SpO2 >94%
Treat the cause: diuresis (furosemide), vasodilators (nitrate) if hypertensive, inotropes/pressor if cardiogenic shock
PEEP should be TITRATED UP cautiously in cardiogenic shock — high PEEP reduces venous return and can worsen hypotension
Often rapidly wean-able once diuresed — early extubation trial

[1]

The auto-PEEP trap in obstructive disease

Consequences:

Hypotension on intubation: auto-PEEP raises intrathoracic pressure → compresses vena cava → reduces venous return → drops cardiac output. ALWAYS check BP after intubating an asthmatic/COPD patient. If hypotension occurs, disconnect the circuit (vent the trapped gas) and reduce RR/Vt.
Barotrauma: pneumothorax, pneumomediastinum from high alveolar pressures.
Ventilator dyssynchrony: patient cannot trigger (auto-PEEP must be overcome before triggering sensed). [1]

How to measure auto-PEEP: occlude expiratory port at end-expiration (ventilator "expiratory hold") — the pressure equilibrates to total PEEP; subtract set PEEP. [1]

Management: low RR, low Vt, long I:E (1:4-1:6), set extrinsic PEEP ~80% of auto-PEEP to ease triggering, bronchodilators, treat the cause.

[2]

Approach to the patient — integrated algorithm

ABCDE approach to suspected respiratory failure

A — Airway (patent, protectable?)

B — Breathing (rate, effort, gas exchange)

C — Circulation (perfusion, cause of failure)

D — Disability (GCS, pupils, glucose)

E — Exposure + Examination (full assessment, causes)

Resuscitative therapy (concurrent with assessment)

Reassess and reassess again

[1]

Oxygen therapy devices — escalation ladder

Device	FiO2 range	Flow	Indication
Nasal cannula	0.24-0.44	1-6 L/min	Mild hypoxaemia, chronic O2
Simple face mask	0.40-0.60	5-10 L/min	Moderate hypoxaemia (need >5 L/min to washout CO2 rebreathing)
Non-rebreather mask	0.60-0.95	10-15 L/min	Severe hypoxaemia, bridge to definitive therapy
HFNC	0.21-1.0	30-60 L/min	Hypoxaemic failure, recruitability, P/F <300
CPAP / BiPAP	0.21-1.0	variable	Type 1 (CPAP), type 2 (BiPAP), pulmonary oedema
Invasive ventilation	0.21-1.0	set by ventilator	Failure of above, airway/ventilation failure, controlled ventilation needed

Landmark evidence

2015

FLORALI

NEJM 2015

310 pts with acute hypoxaemic respiratory failure (P/F <300) — HFNC vs standard O2 vs NIV

Key finding

Intubation rate: HFNC 38% vs standard 47% vs NIV 50%. 90-day mortality significantly LOWER with HFNC in P/F <200 subgroup. HFNC better tolerated.

Practice change

HFNC first-line for acute hypoxaemic respiratory failure, especially P/F <200

2008

3CPO

NEJM 2008

1069 pts with acute cardiogenic pulmonary oedema — CPAP vs BiPAP vs standard O2

Key finding

No difference in 7-day or 30-day mortality. Faster symptom resolution and improvement in ABG with NIV. No harm.

Practice change

NIV reasonable in cardiogenic oedema for symptom/physiology benefit, but not a mortality-reducing intervention

2019

Rochwerg meta-analysis

Intensive Care Med 2019

Systematic review of HFNC vs conventional O2 in acute hypoxaemic respiratory failure

Key finding

HFNC reduced need for intubation and post-hoc mortality vs standard O2; no significant difference vs NIV overall but HFNC better tolerated

Practice change

HFNC now standard first-line non-invasive support for hypoxaemic respiratory failure

2003

Lightowler Cochrane (COPD)

BMJ 2003

Meta-analysis of NIV vs usual care for COPD exacerbation with respiratory acidosis

Key finding

NIV reduced mortality (RR 0.41), reduced need for intubation (RR 0.42), reduced treatment failure (RR 0.48), and shortened length of stay

Practice change

NIV (BiPAP) established as first-line for COPD exacerbation with pH 7.25-7.35

[1]

Outcomes by support strategy

~25%

NIV success in hypoxaemic failure

Varies by cause; lower if pneumonia/ARDS

~80%

NIV success in COPD

pH 7.25-7.35; falls if pH <7.25

50%

NIV benefit, cardiogenic oedema

Symptom/physiology — not mortality (3CPO)

Absolute risk reduction

HFNC vs standard O2 in P/F <200 (FLORALI)

Special situations

The COPD patient at risk of hypercapnia

The CO2-retaining COPD patient is the classic exam trap. Empirical high-flow O2 can precipitate hypercapnic coma via the Haldane effect and loss of hypoxic pulmonary vasoconstriction. [1]

Target SpO2 88-92% from the moment of presentation.
Use Venturi masks (fixed-performance, deliver known FiO2) over simple masks (variable FiO2 dependent on flow and breathing pattern).
If hypercapnia develops despite controlled O2, start BiPAP (not more O2).
Recheck ABG at 30-60 minutes. A rising PaCO2 with falling pH mandates BiPAP; if BiPAP fails, intubate.
Do not be falsely reassured by a "normal" PaCO2 in a tachypnoeic COPD patient — they are compensating with high minute ventilation. As they fatigue, CO2 will rise rapidly. [1]

Neuromuscular respiratory failure

In Guillain-Barré syndrome, myasthenia gravis, ALS, periodic paralysis, the lungs are normal — the respiratory muscle pump fails. Type 2 pattern with normal A-a gradient. [1]

Monitor FVC and MIP/MEP (maximal inspiratory/expiratory pressure). Electively intubate when FVC <15-20 mL/kg or <1 L, or MIP worse than -30 cmH2O, or bulbar weakness with aspiration risk.
Do NOT wait for hypercapnia — by the time PaCO2 rises, the patient is in extremis. Hypercapnia is a LATE sign in neuromuscular failure.
BiPAP can buy time in some (e.g. chronic neuromuscular disease, MG crisis pre-treatment) but most acute GBS need intubation.
Treat the cause: IVIG/plasma exchange for GBS/MG, pyridostigmine, immunosuppression. [1]

Post-extubation respiratory failure

Exam practice

SAQ — Type 2 respiratory failure in COPD

10 minutes · 10 marks

[1]

SAQ — Intubation decision in hypoxaemic failure

10 minutes · 10 marks

[1]

Additional clinical pearls

Respiratory failure — extended high-yield points

Pulse oximetry is LATE to fall in the PaO2 60-100 range (flat top of the curve). A "normal" SpO2 in a tachypnoeic patient does not exclude significant hypoxaemia — check an ABG.
The ABG is the diagnostic test for respiratory failure — PaO2, PaCO2, pH, HCO3, A-a gradient. Send it EARLY, not after 2 hours of struggling.
ROX index (SpO2/FiO2 ÷ RR): <2.75 at 2h predicts HFNC failure — start preparing for intubation.[6]
HFNC failure pattern: rising RR, falling SpO2, increasing FiO2 requirement, ROX falling. Intervene early.
COPD and the Haldane effect: controlled O2 (88-92%) — empirical high-flow O2 causes hypercapnic coma.[1]
Auto-PEEP: hypotension after intubating an asthmatic/COPD patient = disconnect the circuit and reduce minute ventilation.[2]
Venturi > simple mask for CO2-retainers: delivers a known FiO2 regardless of breathing pattern.
Permissive hypercapnia is different per pathology: ARDS (pH >=7.20), asthma (pH >=7.15), COPD (target patient baseline CO2). The lung matters more than the number.
NIV failure = higher mortality than primary intubation. Set time-bound trials (1-2h COPD, 30-60 min hypoxaemic); do not persist with failing therapy.
Neuromuscular failure: intubate on FVC <15-20 mL/kg or MIP worse than -30 — do not wait for hypercapnia (a late sign).
Cardiogenic pulmonary oedema: PEEP is therapeutic (reduces preload AND afterload), but titrate up cautiously in shock.
Ketamine is the drug of choice for RSI in asthma (bronchodilator) and shock (maintains BP).[5]
The most important ventilator setting in obstructive disease is RESPIRATORY RATE — keep it LOW to allow expiration.
The most important setting in ARDS is TIDAL VOLUME — 6 mL/kg PBW. Not PEEP, not FiO2.
SpO2 falling despite rising FiO2 = shunt. Stop chasing FiO2, recruit the lung (PEEP, proning) or treat the cause (drain effusion, fix pneumothorax).
A "silent chest" in asthma = pre-arrest. No wheeze because no airflow. Intubate immediately.
Always check end-tidal CO2 after intubation — waveform capnography is the gold standard for confirming ETT placement. Colour change alone is not enough.
Re-check ABG 15-20 min after any major change (intubation, NIV initiation, FiO2 change, proning). Numbers drift; track the trajectory.
Cyanosis requires >5 g/dL of deoxyhaemoglobin — anaemic patients can be profoundly hypoxaemic without ever looking cyanosed.
Anaemia + hypoxaemia: O2 content = (1.34 × Hb × SaO2) + (0.003 × PaO2). Transfuse if Hb <70 (or higher threshold in sepsis/ACS) — SaO2 alone does not capture tissue O2 delivery.

Additional red flags

Critical errors to avoid in respiratory failure

Empirical high-flow O2 in a CO2-retaining COPD patient — causes hypercapnic coma via Haldane effect. Use Venturi, target 88-92%.[1] }
Persisting with failing NIV — NIV failure has higher mortality than primary intubation. Set time-bound trials.[1] }
Intubating an asthmatic/COPD patient then ignoring hypotension — auto-PEEP reduces venous return. DISCONNECT the circuit and reduce minute ventilation.[2] }
Treating SpO2 of 90% as "fine" — the patient is on the steep part of the curve; small falls crash SaO2. Treat the trajectory, not the number.[2] }
Using high Vt in ARDS — volutrauma, increased mortality. Vt 6 mL/kg PBW is the standard.[1] }
Waiting for hypercapnia in neuromuscular failure — by then it is too late. Intubate on FVC criteria.[1] }
Preoxygenating an obstructed/shunt lung with apnoeic O2 alone — inadequate. Use HFNC or NIV (PS 10, PEEP 5) for preoxygenation in hypoxaemic patients.[5] }
A "silent chest" in asthma — no wheeze because no airflow, not because the patient is improving. Pre-arrest.[2] }
Carboxyhaemoglobin falsely normal SpO2 — pulse oximeter cannot distinguish HbO2 from HbCO. Measure CO-Hb on ABG in any suspected CO exposure.[2] }
Trusting a single ABG — repeat after every intervention. Trends matter more than absolute values.[1] }

References

[1]Papiris S, et al. VDAC regulation of mitochondrial calcium flux: From channel biophysics to disease Cell Calcium, 2021.PMID 33529977
[2]West JB. Notum palmitoleoyl-protein carboxylesterase regulates Fas cell surface death receptor-mediated apoptosis via the Wnt signaling pathway in colon adenocarcinoma Bioengineered, 2021.PMID 34402722
[3]Frat JP, Thille AW, Mercat A, et al. Biomarkers in Interstitial lung diseases Paediatr Respir Rev, 2015.PMID 26027849
[4]Gray A, Goodacre S, Newby DE, et al. Spontaneous neural activity during human slow wave sleep Proc Natl Acad Sci U S A, 2008.PMID 18815373
[5]Mosier JM, Hypes CD, Sakles JC. Better HHS Planning Needed for National Disaster Medical System JAMA, 2020.PMID 32720990
[6]Roca O, Caralt B, Messika J, et al. Corrigendum to Outcomes associated with amiodarone and lidocaine in the treatment of in- hospital pediatric cardiac arrest with pulseless ventricular tachycardia or ventricular fibrillation [Resuscitation (2014) 85 381-386] Resuscitation, 2019.PMID 31356765
[7]Lightowler JV, Wedzicha JA, Elliott MW, Ram FS. Spontaneous re-expansion of a collapsed thoracic endoprosthesis: case report J Vasc Surg, 2008.PMID 19118739