4.3 Airway Management and Tracheostomy

A brain injury does not intubate itself. You intubate because the patient cannot protect the airway, cannot oxygenate or ventilate, or will clearly lose those abilities before the next scan. This independent OpenExamPrep section covers indications, rapid-sequence intubation (RSI) when intracranial pressure (ICP) is high, cervical-spine precautions, tracheostomy indications and timing (including SETPOINT2), and the complications that turn a tube into a new emergency.

Indications to intubate

Classic teaching is Glasgow Coma Scale (GCS) ≤8. That cutoff is a screening rule, not a law of physics. A patient with GCS 7 from bihemispheric injury who still has a gag, a cough, and a patent airway may be watched for minutes while you treat seizures or hypoglycemia. A patient with GCS 11 from a cerebellar hemorrhage who is pooling secretions, gurgling, and rising on PaCO2 needs a tube now. Use GCS with airway reflexes, work of breathing, and trajectory.

Intubate for:

• Loss of airway protection: absent gag/cough, pooled secretions, witnessed aspiration, stertor from a prolapsed tongue or a retropharyngeal hematoma.
• Hypoxemic failure: PaO2 remaining inadequate despite oxygen and noninvasive support, or inability to protect the interface (mask NIV is a poor plan in vomiting ICH).
• Hypercapnic failure: rising PaCO2 with acidosis or declining mental status, including neuromuscular disease (Guillain-Barré FVC falling through 20 mL/kg, myasthenic crisis) and central hypoventilation.
• Anticipated course: malignant MCA edema, planned long operating room, refractory status epilepticus, cervical cord injury with vital-capacity collapse, or a transfer through a CT scanner that will be unsafe if the patient vomits.

Do not intubate solely to “rest the brain” in a protecting, oxygenating patient with a treatable metabolic coma. Do not delay intubation in a herniating patient because the GCS is still 9.

RSI when ICP is high

RSI is simultaneous sedation and paralysis to secure the trachea without bagging a full stomach. In high ICP, the extra constraints are cerebral perfusion pressure (CPP = MAP − ICP) and venous drainage.

Hemodynamics: a propofol induction in a dry TBI patient can drop mean arterial pressure (MAP) 20–40 mm Hg and turn ICP 25 into ischemia. Prefer agents that spare blood pressure when the patient is shocked or volume-down—often ketamine or etomidate—and have vasopressors drawn. Ketamine’s old “raises ICP” reputation is not how a ventilated neuro ICU uses it when PaCO2 is controlled; hypotension is the more common CPP killer. Lidocaine or high-dose opioid pretreatment to blunt the laryngoscopy ICP spike has mixed evidence; do not let a lidocaine ritual delay an obstructed airway.

Laryngoscopy: video laryngoscopy is first-line in most modern neuro ICUs because it shortens the apneic interval and reduces the need to crank the neck. Manual in-line stabilization if trauma is possible. Preoxygenate without crushing cerebral venous return (avoid aggressive Trendelenburg). After the tube is in, ventilate to a normal PaCO2, not to a trophy end-tidal CO2 of 20 “for the brain.” Confirm placement with waveform capnography.

Paralytics: succinylcholine is fast; avoid it after 24 hours in denervating disease, burns, or rhabdomyolysis because of hyperkalemia. Rocuronium 1.2 mg/kg is the usual nondepolarizing RSI dose; remember it erases the examination until it wears off or is reversed (sugammadex). Do not paralyze a patient you have not yet committed to ventilate.

Other caveats: full stomach and ICH often coexist—assume aspiration risk. Avoid nasal tubes in suspected basilar skull fracture (Battle sign, raccoon eyes, CSF leak). Have a difficult-airway plan (supraglottic rescue, cricothyrotomy) before the first look. Head of bed as high as intubation mechanics allow; once the tube is secure, return to 30 degrees and a neutral neck.

Cervical spine

Any unconscious trauma patient, fall, diving injury, or unknown mechanism gets c-spine protection until imaging or a valid clinical clearance says otherwise. That means a collar or equivalent, manual in-line stabilization during laryngoscopy, and video laryngoscopy rather than aggressive sniffing-position extension. Do not remove the collar to “get a better view” without a stabilization plan. Fiberoptic awake intubation is an option in the cooperative unstable-neck patient who is not herniating; it is the wrong delay in a vomiting comatose patient. After intubation, keep the neck neutral—rotation and flexion kink venous outflow and can raise ICP even when the bony spine is intact.

Tracheostomy: why, when, and what goes wrong

Indications include expected prolonged mechanical ventilation, repeated extubation failure, upper-airway obstruction that will not resolve (bilateral vocal-cord injury, massive facial trauma once the acute airway is oral), and secretion management when the patient cannot cough. Severe TBI, brainstem stroke, and high cervical cord injury populate this list because airway protection lags motor recovery.

Timing is not a universal day-7 law. Brain Trauma Foundation material supports early tracheostomy to reduce ventilator days when benefit outweighs procedural risk, without evidence that early tracheostomy reduces mortality or pneumonia. In severe stroke, SETPOINT2 (JAMA 2022) randomized early tracheostomy ≤5 days versus standard weaning with tracheostomy if needed from day 10. Six-month survival without severe disability (mRS 0–4) was 43.5% versus 47.1% (not significant). Early tracheostomy is not a neurorecovery drug. Individualize: a high cord transection will need a surgical airway; a rapidly clearing intoxication should not get a day-3 trach because the GCS was 7 on arrival.

Technique is percutaneous at the bedside or open in the operating room. Coagulopathy, goiter, anomalous vessels, and an extremely short neck raise bleeding and false-passage risk. Anticoagulation, if still indicated, should be held per procedural protocol when the brain will tolerate the hold.

Complications you must name

Cuff leaks: first check that the pilot balloon holds pressure (a ruptured cuff needs a tube change). Next confirm depth—a tube that has migrated so the cuff sits at or above the stoma will leak. Suction above the cuff; pooled secretions mimic a leak. A tube that is too small for the trachea will leak at reasonable cuff pressures—do not keep adding air until you create tracheal ischemia (keep cuff pressure generally around 20–30 cm H2O). A new leak plus subcutaneous air is a false tract until proven otherwise.

Accidental decannulation: if the tract is mature (often after about 7 days, longer if healing is poor), you may recannulate with a smaller tube or obturator, confirming ETCO2. If the tract is immature, the stoma can close in minutes and a forced recannulation creates a false passage into the mediastinum. Orally intubate from above, oxygenate, and call airway expertise. Cover the stoma while bagging so air does not leak out the neck. This is a favorite exam fork: heroic bedside dilation of a 36-hour-old perc tract is how you write a pneumomediastinum note.

Putting airway decisions together

Think in three time scales. Seconds: obstruction, vomiting, apnea—tube now, video laryngoscope, protect the neck, protect MAP. Hours: GCS 7 with a planned hemicraniectomy—intubate before the OR rather than after herniation in the elevator. Days: still intubated, weak cough, thick secretions—discuss tracheostomy without pretending SETPOINT2 promised a better mRS.

Exam traps

Intubating every GCS 8 regardless of airway reflexes and gas exchange—or refusing to intubate a GCS 10 patient who is aspirating. Propofol induction in a hypotensive herniating patient. Nasal intubation through a basilar skull fracture. Cranking the neck in uncleared trauma. Quoting SETPOINT2 as proof that day-3 tracheostomy saves independence. Blindly reinserting a dilator into a fresh perc stoma after decannulation. Overinflating a cuff to “fix” a leak caused by a malpositioned tube. Assuming a tracheostomy means the patient can eat. Using succinylcholine for RSI in a week-old ICU patient with denervating Guillain-Barré.

After the tube is in, write down the induction MAP. A beautiful view that cost 40 mm Hg of systolic pressure is not a successful neurointubation.