Traumatic Brain Injury



Mild injury = GCS 13-15, may be slightly lethargic or confused but can communicate and follow commands. Moderate injury = GCS 9-12, often lethargic or obtunded but arousable to vocal or noxious stimulation. These patients must be evaluated for drug intoxication, hypoxia, hypercarbia, and metabolic abnormalities. Severe injury = GCS 8 or less (coma).

Mechanism of injury

In addition to primary injury, delayed axotomy has been described, which can occur up to 24 hours post-injury [Acta Neuropath (Berl) 89: 537, 1995]. After head injury there is enhanced susceptibility to a second injury, even after mild insult – this has been confirmed in animal models. [Brain Res 477: 211, 1989]

Decision to go to Surgery

The acceptable time initiate non-neurosurgical procedures in the TBI patient is not known. It is known that early surgery is associated with an increased incidence of hypoxemic and hypotensive events [Kalb DC et al. Surgery 124: 739, 1998], and some studies have shown worsened outcomes following intraoperative hypotension [Pietropaoli JA et al. J Trauma 33: 403, 1992 (53 patients, mortality studied); Jaicks RR et al. J Trauma 42: 1, 1997 (33 patients, GCS studied)]. On the other hand, it is known that in some operations, the rate of certain complications (ex. infectious, pulmonary) increase with delayed surgery [Dunham CM et al. J Trauma 50: 958, 2001; Charash WE et al. J Trauma 37: 667, 1994] even in non-thoracic operations. Because of the dangers associated with transport and surgery itself, patients with a severe TBI (GCS < 9) should avoid non-neurosurgical operations whenever possible (i.e., urgent, but not emergent surgery should be avoided).


Airway and Ventilation

Because a single episode of PaO2 < 60 mm Hg is associated with an almost 2-fold increase in mortality (and SBP < 90 mm Hg associated with a 150% increase in mortality) [Chesnut RM et al. J Trauma 34: 216, 1993 (717 cases of GCS < 9, Traumatic Coma Databank)], prehospital intubation and ventilation is indicated in all severe TBI, and if not attained by admission should be an immediate priority.

Positive End Expiratory Pressure

Head injuries in and of themselves do not mandate a particular ventilation strategy, however many of these will be polytrauma patients with chest wall or pulmonary injuries, massive blood loss, and a recent history of aspiration, all of which complicate ventilatory management. Classically PEEP has been avoided because of its potential to increase ICP and decrease cardiac output, however it may be necessary to avoid hypoxemia. One recent study showed that while 10 or more cm H20 increased ICP in patients with normal ICP, it had no effect on those with intracranial hypertension [McGuire G et al. Crit Care Med 25: 1059, 1999 (n = 17)]. Another study of TBI patients (GCS < 8) showed a trend towards decreased ICP and increased CPP when increasing PEEP from 0 to 15 cm H2O. [Huynh T et al. J Trauma 53: 488, 2002 (n = 20)]


Hyperventilation, which decreases cerebral blood flow (through cerebral vasoconstriction, hence the ICP-lowering effect) is inappropriate unless brain herniation is imminent. Additionally, it is impossible to predict whether autoregulation is intact – in the autoregulated brain, hyperventilation (decreased PaCO2) will decrease cerebral blood flow and lower ICP, whereas this may not occur following a TBI [Marion DW et al. Neurosurgery 29: 869, 1991], thus it is theoretically possible to produce hyperemia in the injured brain following hyperventilation. For these reasons, the Brain Trauma Foundation (BTF) guidelines recommend maintaining PaCO2 at 35 mm Hg and lowering to 30 mmHg only for intractable increases in ICP. [Brain Trauma Foundation. J Neurotrauma 17: 452, 2000]


Because a single episode of SBP < 90 mm Hg is associated with a 150% increase in mortality [Chesnut RM et al. J Trauma 34: 216, 1993 (717 cases of GCS < 9, Traumatic Coma Databank)], thus some authors recommend maintaining SBP > 110 mm Hg and MAP > 90 mm Hg (i.e. CPP > 70 assuming ICP is 20), which should provide a margin of safety. The BTF, which is primarily focused on perfusion, recommends maintaining CPP > 70 mm Hg.

Previously TBI patients were fluid-restricted, however most of these patients are under-resuscitated and the effect on hemodynamics was not acceptable. Thus, the current recommendations are to maintain TBI patients in a euvolemic state, which often means administration of fluids. The optimal fluid has not been determined, although hypertonic saline has shown considerable promise.

Selection of Anesthetic / Intraoperative Considerations

No anesthetic regimen has been proven superior, although the combination of low-dose volatile + remifentanil has the advantages of a relatively minimal hemodynamic effects and allowance for rapid emergence/examination. Benzodiazepines should be avoided as they may complicate the post-operative exam. Intraoperatively, one should use the same techniques used in the ICU to manage CPP – position (if possible), hemodynamic monitoring and support if needed, mannitol, and sufficient depth of anesthesia.

Critical Care

Who to admit to the NICU

Any patient with severe TBI, open calvarial, or penetrating injury to the skull. All post-craniotomy patients. Many moderately injured patients, and even some mildly injured patients if they have significant other injuries. Note that 20% of patients with a linear or depressed skull fracture will require operative intervention. [J Neurosurg 65: 203, 1986]


Any 2-point drop in GCS or new focal deficit should be considered significant. Andrews recommends sedating with propofol and avoiding benzodiazepines. Therapeutic paralysis should be a last resort (goal = 1-2 on train of four, not four as is commonly mistaken).

Cardiopulmonary concerns

A-line is mandatory. TLC is recommended. A CVP < 5 mm Hg suggests dehydration in this patient population and should likely be corrected (no data). In elderly patients with ventricular disease, spinal cord injury patients, and patients with ARDS, Andrews recommends a PA catheter and maintenance of wedge 12-14 mm Hg. All severe TBI patients should be intubated. Continuous hyperventilation loses its effect as soon as 4-6 hours [Andrews] but always by 20-24 hours as by this time serum and CSF pH will normalize [Br J Anesth 46: 348, 1974; Scand J Clin Lab Invest 22: 247, 1968] – return of PaCO2 to more normal values then leads to overdilation of arterioles and subsequent elevations in ICP, thus continuous hyperventilation should be avoided. In an experiment involving seven hyperventilated rabbits, vessel diameter was reduced by 13% at 4 hours, 3% at 24 hours, and at 52 hours the vessel diameters were 105% of baseline at a pCO2 of 25 mm Hg and 122% at a pCO2 of 38 mm Hg. [J Neurosurg 69: 923, 1988]

Manual repositioning patients every hour can significantly reduce atelectasis. [Luce JM: Intensive Resp Care, 1984]

When resuscitating neurosurgical patients with other trauma, even in the presence of severe head injury, adequate volume resuscitation is far more important than the fear of excessive fluid overload. [Andrews]

Blood pressure management is controversial – often hypertension is secondary to pain, which should be treated. Other potential causes include cerebral ischemia and elevated ICP – if analgesia does not work, use short-acting beta blockers in most cases to keep SBP < 180 mm Hg [Andrews] – when cerebral ischemia or ICP is an issue, keep SBP < 220 mm Hg but do not treat until then.

Intracranial Pressure in TBI

ICP should probably be monitored in all traumatic brain injury patients with severe injury (GCS 8 or less) and in those with less severe injury but a high index of suspicion. Treatment of ICP > 20 mm Hg [Andrews; J Neurosurg 56: 560, 1981; J Neurosurg 83: 949, 1995] to maintain CPP > 70 mm Hg [J Trauma 30: 933, 1990; J Neurosurg 83: 949, 1995] is advocated by many. Monitor ICP until is has been normal for 24 hours or more.

To manage ICP, remove all potential underlying causes (hematoma). For medical management – raise the head of bed, drain CSF, give temporary doses of IV mannitol and/or hyperosmolar therapy, and sedate with propofol (shown by Class I evidence to be superior to morphine in managing elevated ICP [Bullock R: Guidelines]). Limit free water intake, and consider diuretics, ex. furosemide. The goal is not dehydration but euvolemia – Na 135-145 and mOsm 300-330 are optimal, as is a CVP of 2-5 mm Hg.

Barbiturate coma may be indicated for failure of all other methods, however, based on three available randomized controlled trials, the 2004 Cochrane Database review stated “There is no evidence that barbiturate therapy in patients with acute severe head injury improves outcome. Barbiturate therapy results in a fall in blood pressure in one of four treated patients. The hypotensive effect of barbiturate therapy will offset any ICP lowering effect on cerebral perfusion pressure”.

Hypothermia has been shown to lower ICP [NEJM 336: 540, 1997;J Neurosurg 79: 363, 1993; J Neurosurg 89: 206, 1998] and may improve outcomes in patients with severe head injury [J Neurotrauma 10: 263, 1993; J Neurosurg 79: 354, 1993]. There may be, however significant toxicities associated with hypothermia, thus beware. The Brain Trauma Foundation Guidelines state that there is Class III evidence suggesting that “Prophylactic hypothermia is associated with significantly higher Glasgow Outcome Scale (GOS) scores when compared to scores for normothermic controls.” However, Class III data also suggest that “prophylactic hypothermia is not significantly associated with decreased mortality when compared with normothermic controls.”

Metabolic/Nutrition in the Traumatic Brain Injury

Use 0.9% NaCl without glucose at a rate lower than what you would use for systemic trauma. Glucose has been shown to cause poor outcomes in trauma patients [Neurosurgery 18: 469, 1986; J Trauma 58: 47, 2005]. Keep Na < 150 mEq/L and mOsm < 330 mOsm/L. Early nutrition is thought to be beneficial, however TPN in TBI patients who can’t eat has been shown to improve nitrogen balance but does not affect survival. [Neurosurgery 19: 367, 1986]

Miscellanei: Traumatic Brain Injury

  • Only use 0.9% NaCl without glucose and tight glucose control
  • Keep Na < 150 mEq/L and Osm < 330 mOsm/L
  • TPN does not affect mortality
  • Phenytoin is superior to valproate, magnesium, and steroids for early seizure prophylaxis


Load Phenytoin 15-18 mg/kg with 5 mg/kg after. Temkin’s study of 404 TBI patients given dilantin within 24 hours of injury and showed that a year of therapy was no more effective than seven days (short term incidence reduced from 13.3 to 3.4% [NEJM 323: 497, 1990]) – no other therapy has proven superior. Valproate shows no benefit and may increase mortality [J Neurosurg 91: 593, 1999], magnesium may be detrimental [Lancet Neurol 6: 29, 2007], and steroids do not work. [Epilepsia 45: 690, 2004]