Airway and RSI

In depth

The one-page summary above is what the PDF prints. These sections go further.

The physiologically difficult airway: resuscitate first

In the ICU most peri-intubation harm is physiological, not anatomical. In INTUBE (Russotto, JAMA 2021; 2,964 patients in 29 countries), 45.2% had at least one major adverse peri-intubation event: cardiovascular instability 42.6%, severe hypoxemia 9.3%, cardiac arrest 3.1%. Induction removes sympathetic tone, positive pressure cuts venous return, and apnea stops CO₂ clearance while a shunted lung desaturates within seconds. Mosier's "HOp killers" (hypotension, oxygenation, pH) name the states to fix first; right ventricular failure is the fourth.

Fluids alone do not prevent collapse: in PREPARE II (Russell, JAMA 2022; 1,065 patients) a 500 mL bolus at induction did not reduce cardiovascular collapse (21.0% vs 18.2%). For most patients the more reliable step is a vasopressor started before induction. Push-dose phenylephrine (50–200 mcg) or epinephrine (10–20 mcg) covers the gap; label every syringe, because dilution errors are the main hazard.

In severe metabolic acidosis (DKA, salicylate toxicity, lactic acidosis) the patient may be moving 20–30 L/min to hold a PaCO₂ near 15–20 mm Hg. A default ventilator rate of 14–16 lets the pH fall further within minutes; set the rate around 30 and check a gas within 30 minutes. In RV failure the spiral is low MAP, RV ischemia, lower output and lower MAP: keep MAP up before induction, start an inhaled pulmonary vasodilator if available, avoid hypoxemia and hypercapnia, and use modest PEEP.

Preoxygenation and apneic oxygenation

Preoxygenation replaces the nitrogen in the functional residual capacity with oxygen; the target is an end-tidal O₂ of at least 85–90%. In Benumof's modeling a healthy, fully preoxygenated adult takes about 6–8 minutes to reach SpO₂ 90%; an obese or critically ill patient with shunt can get there in under 2. Head up 20–30° or ramped (ear level with the sternal notch) enlarges the FRC.

PREOXI (Gibbs, NEJM 2024; 1,301 ED and ICU patients) compared NIV with an oxygen mask: hypoxemia (SpO₂ under 85%) fell from 18.5% to 9.1%, cardiac arrest was 0.2% vs 1.1%, and aspiration did not increase (0.9% vs 1.4%). PreVent (Casey, NEJM 2019; 401 ICU patients) found gentle bag-mask ventilation between induction and laryngoscopy cut severe hypoxemia from 22.8% to 10.9% without more aspiration; older exam items may still expect "do not bag after paralysis".

Apneic oxygenation works because the alveoli take up about 250 mL/min of oxygen and return little CO₂, so gas flows toward them through an open airway; CO₂ still rises. The randomized trials are neutral: FELLOW (Semler, AJRCCM 2016; ICU) found lowest SpO₂ 92% vs 90%, and ENDAO (Caputo, Acad Emerg Med 2017; 200 ED patients, cannula at 15 L/min or more) 92% vs 93%. Observational data favor it and it is cheap, so most protocols keep it.

Induction and paralysis: pharmacology in depth

Give the induction agent and then the paralytic at once, each as a rapid push with a flush. In shock, cut the induction dose by about half (a smaller central volume gives higher brain and heart levels) and raise rocuronium toward 1.2–1.6 mg/kg, because slow circulation delays onset. Dose succinylcholine by total body weight; many references dose rocuronium by ideal or lean weight in obesity.

Ketamine's direct effect on the myocardium is depressant, which shows when catecholamine stores are exhausted. That is the likely reason the RSI trial (Casey, NEJM 2026; 2,364 patients) found more cardiovascular collapse with ketamine than etomidate (22.1% vs 17.0%) and no difference in 28-day mortality. Etomidate suppresses cortisol synthesis for 6–8 hours; SCCM 2023 found no mortality difference and does not recommend routine steroids after one dose.

Succinylcholine depolarizes the acetylcholine receptor, causing fasciculations and then flaccid paralysis, and is hydrolyzed by plasma cholinesterase. Potassium rises about 0.5–1 mEq/L normally. After denervation, burns, crush, prolonged immobility or myopathy, extrajunctional receptors spread over the muscle and the rise can arrest the heart; the risk starts after about the first day and peaks around 7–10 days. It also triggers malignant hyperthermia and masseter spasm and causes bradycardia with repeat doses and in children.

Rocuronium is a competitive nondepolarizing blocker cleared by the liver and bile, so it lasts longer in liver failure and in older patients. The Cochrane review (Tran 2015) found succinylcholine superior overall for excellent intubating conditions but no significant difference at rocuronium 1.2 mg/kg. The short action of succinylcholine is rarely a rescue in the ICU: a critically ill patient desaturates long before breathing returns. Sugammadex 16 mg/kg reverses rocuronium within about 3 minutes but does not open an obstructed upper airway.

Predicting difficulty, choosing the device, and the failed airway

Video laryngoscopy is now the default in critically ill adults. DEVICE (Prekker, NEJM 2023; 1,417 ED and ICU patients) found first-attempt success of 85.1% with video vs 70.8% with direct laryngoscopy, with similar severe complications; it stopped early for efficacy. A hyperangulated blade gives an easy view but a harder pass and needs a rigid stylet shaped to the blade.

The bougie helps most with a direct laryngoscope and a poor view. In BEAM (Driver, JAMA 2018; 757 ED patients, mostly direct) first-attempt success was 98% vs 87%, and 96% vs 82% with a difficult airway feature. In BOUGIE (Driver, JAMA 2021; 1,102 ED and ICU patients, mostly video) it was 80.4% vs 83.0% with a stylet, no difference, and severe hypoxemia was 11.0% vs 8.8%. Keep it ready as a rescue rather than routine with video.

Each attempt costs oxygen and swells the airway; Mort (Anesth Analg 2004) found more than two attempts multiplied hypoxemia, aspiration and arrest several-fold. The DAS critically ill guideline (Higgs, BJA 2018) limits Plan A to 3 laryngoscopy attempts plus 1 by an expert, then Plans B and C, oxygenation with a second-generation supraglottic airway (up to 3 attempts) or two-person face mask, then Plan D, front-of-neck access. The Vortex approach (Chrimes) gives each of the three lifelines (face mask, SGA, tube) one best effort of up to 3 attempts; capnography-confirmed oxygenation by any of them is the green zone, and failure of all three is can't intubate, can't oxygenate (CICO).

1. Declare CICO out loud; give a full paralytic dose; keep oxygen on face and SGA.
2. Extend the neck; find the cricothyroid membrane with the laryngeal handshake.
3. Transverse stab through skin and membrane with a size 10 blade, cutting edge toward the operator.
4. Rotate the blade 90° so the sharp edge points toward the feet; pull it toward the operator.
5. Slide a bougie along the blade into the trachea; railroad a 6.0 cuffed tube; inflate and confirm CO₂.
6. Membrane not palpable: 8–10 cm vertical midline incision, blunt dissection, then the same steps.

Confirmation and the first hour after intubation

Waveform capnography confirms the trachea (sustained trace for at least 7 breaths, PUMA 2022); in cardiac arrest the values are low but the waveform persists. The chest x-ray confirms depth, not placement: tip about 3–5 cm above the carina with the head neutral. Neck flexion moves the tip toward the carina and extension moves it up, roughly 2 cm each way; a deep tube usually enters the more vertical right mainstem.

Rocuronium lasts about an hour; etomidate wears off in minutes. Start analgesia (usually fentanyl) and then sedation within minutes of confirmation; ED-AWARENESS (Pappal, Ann Emerg Med 2021) linked awareness with paralysis to delayed sedation after long-acting paralytics. Once the paralytic wears off, PADIS 2018 favors light sedation (RASS −2 to +1) with propofol or dexmedetomidine over benzodiazepines. Then secure the tube, record depth, place an orogastric tube, raise the head of the bed, check cuff pressure and get an x-ray and a blood gas.

Cuff and tube care, and unplanned extubation

Keep the cuff at 20–30 cm H₂O, measured with a manometer, not by feel. Above about 30 cm H₂O it exceeds tracheal mucosal perfusion pressure (Seegobin 1984), risking ischemia, stenosis, tracheomalacia and tracheoesophageal fistula; below 20, secretions pooled above it leak into the lungs. Recheck each shift, after moving the tube or patient, and during air transport, because cuff volume expands with altitude.

The SHEA/IDSA/APIC 2022 compendium lists as essential: avoid intubation where NIV or high-flow works, minimize sedation, daily paired awakening and breathing trials, early mobility, head of bed 30–45°, and subglottic secretion drainage tubes for an expected stay over 48–72 hours. It recommends toothbrushing and no longer recommends routine chlorhexidine oral care; some units still use it.

• Document depth at the teeth or lips each shift and after every repositioning.
• Turn and transport with one person whose only job is the tube; take circuit weight off the tube.
• Use a bite block if the patient bites; suction above the cuff before deflating it.

Unplanned extubation is mostly self-extubation, the rest accidental displacement during care. Risk factors are agitation and delirium, light sedation without close observation, weak securement, turning and transport, and the weaning phase; restraints do not reliably prevent it. Many self-extubations occur in patients close to weaning and a sizable share never need reintubation; accidental extubations more often do.

1. Call for help and stay with the patient.
2. Tube partly out: hold it, do not push it back blindly; check capnography.
3. No CO₂ or tube out: deflate the cuff, remove it, suction, give high-flow oxygen by mask or bag-mask.
4. Assess work of breathing, SpO₂ and mental status; prepare for a possibly harder reintubation.

Extubation readiness and post-extubation stridor

Pair a daily spontaneous awakening trial with a spontaneous breathing trial: in the ABC trial (Girard, Lancet 2008) this gave more ventilator-free days and lower 1-year mortality (number needed to treat about 7). An SBT lasts 30–120 minutes on pressure support 5–8 with PEEP 5 or on a T-piece. Stop it for RR over 35, SpO₂ under 90%, HR over 140 or a 20% change, SBP over 180 or under 90, agitation, diaphoresis or accessory muscle use.

ATS/CHEST 2017 recommends preventive NIV right after extubation in high-risk patients (for example age over 65, heart failure, COPD, hypercapnia). HIGH-WEAN (Thille, JAMA 2019) found high-flow alternating with NIV cut reintubation to 11.8% from 18.2% with high-flow alone; in low-risk patients high-flow cut it to 4.9% from 12.2% with conventional oxygen (Hernández, JAMA 2016). Do not use NIV to rescue established failure: Esteban (NEJM 2004) found higher ICU mortality, attributed to delayed reintubation.

Post-extubation stridor comes from laryngeal edema; risk rises with traumatic or repeated intubation, intubation over about 6 days, a large tube and female sex. For the cuff leak test, deflate the cuff on volume control and compare inspired with expired volume; a leak under 110 mL (Miller and Cole 1996) or under about 10–25% of tidal volume predicts stridor. ATS/CHEST 2017 suggests the test in high-risk patients (conditional, very low certainty) and, if it fails, systemic steroids at least 4 hours before extubation without retesting. François (Lancet 2007) gave methylprednisolone 20 mg 12 hours before and every 4 hours until extubation; laryngeal edema fell from 22% to 3%.

• Sit upright; humidified oxygen; calm the patient (anxiety raises turbulent flow).
• Nebulized epinephrine; watch for rebound as it wears off; IV steroid if not given.
• Heliox (70:30 or 80:20) lowers turbulent resistance as a bridge.
• Rising work of breathing: reintubate early with a smaller tube.

Tracheostomy care and emergencies

Tracheostomy allows lighter sedation, easier weaning, oral care and communication, but timing does not change survival: TracMan (Young, JAMA 2013) found 30-day mortality of 30.8% with early (day 1–4) vs 31.5% with late (day 10 or later) tracheostomy. The tract takes about a week to mature, so a tube displaced in the first week can be reinserted into pretracheal tissue (false passage).

• At the bedside: obturator, spare tubes in the same size and one smaller, syringe, suction, bag-valve-mask and the bedhead sign.
• Humidify; clean or change the inner cannula per the tube type.
• Change ties with two people, one holding the flange; one finger fits under the ties.
• Cuff 20–30 cm H₂O; suction to a measured depth, under about 15 seconds per pass.

The National Tracheostomy Safety Project (McGrath, Anaesthesia 2012) separates a tracheostomy with a patent upper airway (green bedhead sign) from a laryngectomy, where the mouth does not connect to the lungs (red sign). For a blocked or displaced tube:

1. Call the airway expert; check the bedhead sign.
2. Look, listen and feel at mouth and tube; put waveform capnography on the tube.
3. High-flow oxygen to face and stoma (stoma only after laryngectomy).
4. Remove the speaking valve or cap and the inner cannula.
5. Pass a suction catheter: if it passes, suction and reassess.
6. If it does not pass, deflate the cuff; no improvement, remove the tube. Pulseless: CPR.
7. Oxygenate: bag-mask over the mouth with the stoma covered, or a small mask or SGA over the stoma.
8. Expert: oral intubation with an uncut tube past the stoma, or a small cuffed tube into the stoma.

Early bleeding is usually wound oozing, managed with pressure, packing and correcting coagulopathy. Tracheo-innominate artery fistula is rare (about 0.1–1%), occurs mostly within 3–4 weeks and is often preceded by a small sentinel bleed. Overinflate the cuff; if bleeding continues, press a finger through the stoma against the artery and sternum (Utley maneuver), secure the airway from above with the cuff below the bleed, and call surgery.

Pediatric differences

Doses are weight based: rocuronium 1 mg/kg, succinylcholine 2 mg/kg in infants and 1–1.5 mg/kg in older children (boxed warning for hyperkalemic arrest in undiagnosed myopathy), ketamine 1–2 mg/kg, etomidate 0.3 mg/kg, and atropine 0.02 mg/kg with no minimum may be considered before emergency intubation (AHA/AAP 2025). The cricothyroid membrane is tiny, so surgical cricothyrotomy is generally avoided under about 8–12 years (cutoffs vary) in favor of needle cricothyrotomy.

Exam pearls

• LEMON predicts hard laryngoscopy, MOANS hard bag-mask, RODS a hard SGA, SHORT a hard cricothyrotomy.
• Fix hypotension, hypoxemia and acidosis before induction; set a high rate after intubating a patient in metabolic acidosis.
• Hypoxemic patient: preoxygenate with NIV; apneic oxygenation needs an open airway.
• Etomidate is the most hemodynamically neutral; propofol causes the most hypotension; ketamine bronchodilates.
• Succinylcholine raises potassium about 0.5–1 mEq/L; avoid in hyperkalemia, MH risk, myopathy, and burns or crush after the first day.
• Capnography confirms the trachea; the x-ray confirms depth (tip 3–5 cm above the carina).
• Cuff 20–30 cm H₂O; failed cuff leak test in a high-risk patient means steroids at least 4 hours before extubation.
• Speaking valve only with the cuff fully deflated.
• Blocked tracheostomy: oxygen to face and stoma, remove the inner cannula, pass a suction catheter, deflate the cuff, then remove the tube.