CBF and CPP

Pressure autoregulation is a fundamental concept in cerebrovascular physiology and is distinct from metabolic autoregulation. Pressure autoregulation allows a constant supply of oxygen and metabolic substrates to be delivered to the brain by maintaining a constant CBF over a range of CPPs. CPP is defined as the mean arterial pressure (MAP) minus the intracranial pressure (ICP). Under normal circumstances, CBF remains relatively constant over a range of CPPs between 50 and 150 mm Hg, ensuring adequate oxygen delivery through a dynamic system of arteriolar vasoconstriction and dilation. Hence, in situations of low perfusion pressures, arteriolar vasodilation occurs, resulting in decreased cerebral vascular resistance and, in turn, increased CBF to normal. On the contrary, when a state of high perfusion pressure exists, arteriolar constriction leads to a decrease in CBF toward normal levels. At CPPs outside the autoregulatory threshold, however, CBF is unable to be maintained within its normal parameters—an exceedingly low perfusion pressure results in arteriolar collapse and a consequent decrement in CBF; hyperemia occurs at high CPPs outside of the limit of autoregulation due to vessels reaching their capacity for vasoconstriction (Fig. 132-1).

FIGURE 132-1 Cerebral pressure autoregulation.

(From White H, Venkatesh B. Cerebral perfusion pressure in neurotrauma: a review. Anesth Analg 107:979-988, 2008.)

The surgeon must also be able to distinguish CBF from cerebral blood volume (CBV). As already mentioned, CBF is the physiologic parameter that governs cerebral perfusion and oxygen delivery. CBV represents the total volume of blood within the intracranial compartment and is a key component of ICP per the Monro-Kellie doctrine. In efforts to control ICP, CBV may be manipulated while maintaining an adequate degree of CBF.

Low CPP has been shown to be associated with higher mortality rates and increased morbidity among survivors following TBI,^28–30 especially when systemic hypotension occurs concomitantly.³¹ Moreover, the normal response of cerebral arterioles to changes in CPP is frequently absent or impaired in 49% to 87% of patients after severe TBI.^32–34 Aberrant pressure autoregulation at a range outside normal physiologic parameters may occur after TBI (Fig. 132-2),^35,36 with significant consequences. Shifting the lower limit of autoregulation from 50 mm Hg to a supranormal range of 70 to 90 mm Hg could precipitate cerebral ischemia,^29,37,38 which is especially noteworthy when considering that up to 60% of patients with severe brain injury experience a reduction in CBF during the first few hours following their trauma.³⁹ Prolonged systemic hypotension or aggressive hyperventilation therapy would further exacerbate ischemia in vulnerable areas and could lead to subsequent cerebral edema. Conversely, autoregulation occurring at a lower-than-typical range of CPPs could result in malignant hyperemia or significant damage to the microcirculation, leading to secondary hemorrhages.⁴⁰ Thus, absent or impaired cerebral pressure autoregulation represents a particularly significant risk factor for development of secondary brain injury, especially early in the postinjury period.⁴¹

FIGURE 132-2 Aberrant cerebral pressure autoregulation after severe TBI.

(From Rangel-Castilla L, Gasco J, Nauta HJ, et al. Cerebral pressure autoregulation in traumatic brain injury. Neurosurg Focus 25:1-8, 2008.)

Based on the preceding information, CBF is assumed to have a linear relationship with CPP in managing patients with severe TBI. Other key factors that affect CBF include local factors such as physical compression of vessels by mass lesions or edematous brain, reduced cerebral metabolism of oxygen, and post-traumatic vasospasm.^42,43 Nearly 33% of severely head-injured patients have CBF values near the ischemic threshold (less than 18 ml/dl) within the first 24 hours after primary injury,^44,45 which is seen most often in individuals with diffuse cerebral edema or acute subdural hematomas.^46–48 Low CBF values during the first 7 days after injury are associated with significantly higher mortality rates and increased morbidity.^{32,39,49–51}

CPP thus serves as an important gauge of the adequacy of CBF and, in turn, must be monitored and maintained in patients with severe TBI. Through studies of brain tissue oxygen tension (P_btO₂) and S_jvO₂ that correlate with unfavorable outcomes, some investigators have suggested a treatment and maintenance threshold CPP of 60 mm Hg,^52,53 while other groups recommend a higher value of 70 mm Hg^54,55 to avoid cerebral oxygen desaturation. Based on the results of cerebral microdialysis monitoring, others have discussed a lower CPP threshold of 50 mm Hg.⁵⁶ In a prospective study in which a CPP-focused management strategy was used, Rosner et al.⁵⁷ demonstrated a mean mortality rate of 21% in patients with severe TBI when CPP is maintained above 70 mm Hg, as compared to the 40% mortality rate in the TCDB series. Current practice guidelines dictate a threshold around 60 mm Hg due to concern over causing secondary brain injury in patients with aberrant pressure autoregulation, with further adjustments based on ancillary neurophysiologic monitoring of CBF, oxygenation, metabolism and the individual patient’s pressure autoregulatory status.⁵⁸

Systemic Contributors to Ischemia: Hypotension and Hypoxia

Systemic hypotension and hypoxia have deleterious consequences on patient outcomes following TBI by contributing to secondary brain injury. Episodes of hypotension—defined as systolic blood pressure less than 90 mm Hg; the number of hypotensive events, inclusive of prehospital hypotensive events;^13,59 and the increased total duration of such episodes—increase patients’ mortality and morbidity,⁶⁰ with one study showing a 150% increased risk of mortality in patients with severe TBI who experienced at least one episode of hypotension.⁶¹

Similarly, hypoxia portends poor outcomes after TBI. Stochetti et al.⁶² noted that among severely brain-injured patients, those who had documented oxygen saturations of less than 60% had a mortality rate of 50%, with all survivors being severely disabled, in contrast to a mortality rate of 14.3% and a 4.8% rate of severe disability among nonhypoxic patients. As with hypotension, the number of hypoxic events, along with the duration of hypoxia,⁶⁰ is predictive of worse outcomes. Establishing an airway early in the postinjury period is one important mechanism to prevent hypoxia; early intubation improves outcomes in patients with severe head injuries.⁶³

Brain Oxygen Monitoring

Brain oxygen monitoring is a means of directly assessing the adequacy of oxygen delivery to brain tissue. Commonly employed modalities toward this goal include S_jvO₂ sampling via an internal jugular vein catheter directed toward the jugular bulb and direct brain tissue oxygen monitoring by a fiberoptic catheter. Episodes of jugular venous oxygen desaturation (S_jvO₂ less than 50%) are known to be associated with increased mortality and worse outcomes in patients after severe TBI.^26,64 Low values of P_btO₂ (less than 10-15 mm Hg) and the extent of their duration (more than 30 min) are also associated with high rates of mortality.^65–67 The current guidelines for severe TBI management advocate treatment thresholds for S_jvO₂ and brain tissue oxygen tension at 50% and 15 mm Hg, respectively.⁶⁸

Assessment of CBF via Other Modalities

Modalities aimed at directly measuring CBF include placement of thermal diffusion probes and noninvasive techniques such as transcranial Doppler sonography, xenon CT, nitrous oxide uptake, and positron emission tomography. Assessing the metabolic state of the brain via cerebral microdialysis gauges the adequacy of CBF.⁶⁸ Finally, laser Doppler flowmetry and laser flow flowmetry, performed during surgery with the brain exposed, represent other technologies that measure CBF.^69,70

ICP Monitoring Devices, Indications, and Techniques of Usage

The Association for the Advancement of Medical Instrumentation has developed the American National Standard for Intracranial Pressure Monitoring Devices, which dictates performance requirements of currently available ICP monitoring devices. These parameters include the ability to (1) measure ICP readings from 0 to 100 mm Hg, (2) be accurate within 2 mm Hg in pressure ranges of 0 to 20 mm Hg, and (3) have a maximum error of 10% in pressure ranges of 20 to 100 mm Hg. Current technology allows for ICP measurement by external strain gauge transducers, microtip catheter strain gauge transducers, fiberoptic transducers, and coupling to a pneumatic transducer. There are many possible intracranial locations in which a particular ICP monitoring device may be placed, including within the ventricle, the brain parenchyma, the subdural space, the subarachnoid space, and epidurally. When considering which device to use, attention must be paid to accuracy, reliability, cost, and potential for patient morbidity. A ventricular catheter coupled to an external strain gauge transducer by fluid coupling is the gold standard.⁹⁸

In the setting of severe TBI, placement of an external ventricular drain (EVD), or ventriculostomy, for ICP monitoring also allows for CSF diversion and is useful for patients believed to be at risk for development of intracranial hypertension or hydrocephalus. As such, it has a crucial role in guiding therapy, as well as serving as a therapeutic modality by reducing ICP on a sustained basis.⁹⁹ The most common complications associated with the procedure are hemorrhage, infection, and malfunction. A recent meta-analysis estimated the risk of hemorrhage due to EVD placement as 5.7%, with only 0.61% of hemorrhages being clinically significant¹⁰⁰; however, the significantly higher overall hemorrhage rates reported by other investigators highlight that obtaining postprocedural head CT scans may be a confounding factor.¹⁰¹ Lozier et al.’s¹⁰² review of ventriculostomy-associated infections (VAIs) cited a rate between 6% and 9%. Other important considerations involving VAIs include the significant association of CSF infections with increasing duration of EVD catheterization, the utility of administering prophylactic antibiotics prior to EVD placement, and the use of antimicrobial-impregnated catheters, which have been shown to reduce the risk of EVD-related infections.¹⁰³ Based on the previously cited review of VAIs, prophylactic catheter exchange to reduce VAI rates is not efficacious. EVD malfunctions are usually attributable to obstruction in the fluid coupling mechanism by which the external strain gauge catheter transduces ICP, a complication estimated to occur in 2.5% of patients.¹⁰⁴

At our institution, right-sided EVDs are preferred to avoid the dominant hemisphere. The patient is maintained in a supine position with the head brought to the edge of the bed; the head of the bed is elevated to roughly 45 degrees. Hair clippers are used to widely shave the head on the same side of the EVD cannulation, and the head is prepared in the standard sterile fashion. Two lines are drawn on the scalp to establish Kocher’s point: one line in the midline from the nasion to a point 10 cm back and another from the previous point to a spot 3 cm lateral to it. This location is where the bur hole placement will occur and should be in the ipsilateral midpupillary line. Finally, two lines are drawn that guide the trajectory of ventricular cannulation: one line from Kocher’s point to the ipsilateral medial canthus of the eye and a second line from Kocher’s point to a point roughly 1.5 cm anterior to the ipsilateral tragus (Fig. 132-6). After administration of a local anesthetic containing epinephrine, a small linear skin incision is made down to bone and the periosteum is scraped off. A twist drill is used to open the cranium, bone fragments and dust are removed, and the dura and pia mater are pierced with a scalpel or sharp object. Using the demarcated lines as a guide, the trajectory at which the drill is used to create the bur hole is the same trajectory used for passing the ventricular catheter: aiming in the coronal plane toward the medial canthus of the ipsilateral eye and in the anteroposterior plane toward a point 1.5 cm anterior to the ipsilateral tragus. This trajectory is exceedingly important because it delineates the trajectory of the ventricular cannulation; if the bur hole is not created in an appropriate trajectory, the EVD cannot be optimally positioned due to its being limited by its bony opening, especially if the cranium is thick (Fig. 132-7). The ventricular catheter is passed to a depth of 6.25 cm using the previously described trajectory. The surgeon may alternatively pass the ventriculostomy using Dandy’s principle, which states that employing a trajectory directly perpendicular to the skull allows ventricular cannulation. Regardless, the ventricular catheter is never passed to a depth greater than 7 cm due to the risk of damaging the brain stem. Once the EVD is in place, it is tunneled through the skin away from the incision through a separate stab incision and is secured in place. A postprocedural head CT is then obtained.

FIGURE 132-6 Ventriculostomy anatomy.

FIGURE 132-7 Suboptimal EVD trajectory and placement.

There are cases in which ventricular cannulation can challenge even the most experienced of neurosurgeons. Our experience is to always use the landmarks and trajectory as outlined earlier, regardless of whether a mass lesion or significant shift is present. Most often, errant EVD passes are due to a lateral trajectory (Fig. 132-8), and the neurosurgeon must pay particular attention to avoiding an injury to the ipsilateral internal capsule. In cases in which the ventricle is believed to be cannulated with no egress of CSF, the neurosurgeon may consider attaching a small, plungerless syringe filled with sterile saline to the ventriculostomy catheter. If the ventricle is indeed cannulated, saline should flow from the syringe, through the catheter, and into the ventricle. If the neurosurgeon does not obtain CSF flow through the ventriculostomy after 3 passes, the EVD is left in place and a head CT scan is obtained to determine its position.

FIGURE 132-8 Errant EVD passes.

Other ICP monitoring devices are frequently used in centers, sometimes as an adjunctive means of monitoring ICP when they are placed in tandem with an EVD and at other times when slit-like ventricles preclude EVD placement. The scalp incision is made in the exact location as for a ventriculostomy, with a smaller drill bit and small stylet used to puncture the dura. When using microtip strain gauge or fiberoptic parenchymal monitors, the probe is advanced between 1 and 2 cm into the brain parenchyma. The microtip strain gauge probe must be tunneled through the skin via a separate stab incision and then be secured in place. However, the fiberoptic monitor can be held in place by a fixation bolt screwed into the skull. The reduced accuracy and reliability, increased propensity for malfunction, and high cost in comparison to EVD placement are factors that must be considered prior to using these alternative ICP monitors.

Positioning

Maintaining the head in a neutral position with elevation of the head of the bed to 30 degrees decreases the intracranial blood volume by maximizing venous return from the brain while preserving CPP and CBF.^108–110 ICP reduction may thus be readily achieved without increasing the risk of cerebral ischemia.

Cervical Collar Removal

Of pertinence to patients with severe brain injuries is cervical spine clearance, and subsequent cervical collar removal, on ICP reduction. Studies have shown that patients involved in blunt trauma with a GCS score of 3 to 8 points have a higher incidence of cervical spine injuries as compared to other trauma patients,^111,112 with rates of 10% to 14%.^113,114 Recent investigations have demonstrated the efficacy of a negative helical cervical CT scan in cervical spine clearance among obtunded trauma patients without neurologic deficits that could be potentially explained by spinal cord pathology.^115–118 Therefore, cervical spine magnetic resonance imaging (MRI) is unnecessary to remove external cervical orthoses designed for cervical immobilization, which are known to cause a significant and sustained elevation in baseline ICP.^119–122 Hence, our institutional policy is that among patients with severe TBIs and no neurologic deficits that can be associated with injury to the spinal cord, cervical collars are expediently removed following a negative helical CT scan of the cervical spine.

Maintenance of Normocapnia

As previously mentioned, alterations in arterial PCO₂ result in variations in arteriolar caliber with a consequent effect on CBF and ICP based on the phenomenon of CO₂ reactivity. The physiologic mechanism by which the caliber of cerebral arterioles varies involves changes in the pH of perivascular tissue secondary to P_aCO₂ levels.²⁷ In turn, hypoventilation increases P_aCO₂ and ICP, whereas hyperventilation reduces ICP at the expense of CBF. Maintenance of normocapnia (P_aCO₂ between 35 and 40 mm Hg) provides a balance between ensuring adequate CBF to avoid ischemia and preventing the development of intracranial hypertension associated with hypoventilation.

Sedation

A number of pharmacologic agents are used in patients with severe TBI, with the goal of treating pain and agitation and minimizing the effect of a multitude of noxious stimuli, which may contribute to elevations in blood pressure, body temperature, and ICP and resistance to controlled ventilation.¹²³ In addition to using sedative agents in these patients, the surgeon often needs to determine their efficacy and titrate accordingly. Bispectral index monitoring is a useful tool to accomplish this, because it can provide objective information as to the degree of sedation in these critically injured patients.¹²⁴ The effects of commonly used analgesic and sedative agents on MAP, CBF, and ICP are shown in Table 132-4.

Propofol, a short-acting sedative anesthetic with a rapid onset of activity, is frequently used and has been shown to have superior effects on ICP reduction compared to narcotic-based agents.¹²⁵ In addition, studies have demonstrated that propofol decreases cerebral metabolism and oxygen consumption, has anti-inflammatory properties, and is an antioxidant, leading to the possibility of its having a neuroprotective effect.^126,127

Narcotics have also commonly been employed for providing sedation and analgesia, both alone and with other agents. Morphine sulfate has been shown to provide excellent analgesia but has poor sedative properties. Tachyphylaxis may be seen that can lead to a need for dose escalation and a potential for withdrawal when therapy is discontinued.¹²³ Rapidly metabolized synthetic narcotics such as fentanyl, sufentanyl, and remifentanil are being used more often due to their brief duration of action. However, increases in ICP and decreases in MAP seen with fentanyl and sufentanyl usage mitigate their utility.^128,129

Benzodiazepines have the untoward side effect profile of being neurodepressants and exhibiting long half-lives and thus are not considered ideal sedative agents in the brain-injured population. Midazolam, however, is a commonly employed agent in the management of patients with head injuries due to its short half-life compared to other benzodiazepines and its relatively minimal effect on hemodynamics, especially in comparison to propofol. Midazolam exhibits a similar safety and efficacy profile versus propofol,^130,131 with a decrement in the cerebral metabolism and oxygen consumption seen with benzodiazepine administration being a potential additional benefit.

Dexmedetomidine, an alpha-2 agonist, is one of the newer agents to be studied and used in neurosurgical patients. Increased CPP and decreased ICP have been seen with its use, which, combined with a benign pharmokinetic profile,¹³² warrants further study for its use in patients with severe TBI.

Hyperosmolar Therapy

Neuromuscular Paralysis

The physiologic rationale for using paralytic agents in the treatment of elevated ICP is predicated on their ability to reduce intrathoracic pressure, thus maximizing venous outflow and decreasing ICP. Pharmacologic paralysis prevents any muscle activity, in the form of posturing movements, shivering, or “bucking” the ventilator, from raising the respiratory muscle tone and triggering an increase in intrathoracic pressure.¹⁵⁸ Hsiang et al.¹⁵⁹ reviewed data from the TCDB and concluded that neuromuscular paralysis should only be employed for the management of refractory intracranial hypertension in patients with severe TBI. Early, routine, long-term use of such agents did not improve outcome and led to longer intensive care unit stays and higher rates of infectious complications, such as pneumonia and potentially sepsis.

If a neuromuscular blockade must be employed to address refractory intracranial hypertension in a patient with severe TBI, agents with a short half-life (e.g., vecuronium) are preferable since long-lasting pharmacotherapeutics preclude the patient’s participation in a neurologic examination.^160,161

Hyperventilation

Hyperventilation therapy for the management of intracranial hypertension was introduced in the 1950s by Lundberg, a Swedish neurosurgeon, and his colleagues.¹⁶² The premise behind hyperventilating patients with high ICPs is based on the responsivity of arteriolar smooth muscle in the cerebral vasculature to changes in arterial CO₂. CBF changes 2% to 3% for each millimeter of mercury change in P_aCO₂; thus, hypercarbia from hypoventilation causes reactive vasodilation and increased CBF, whereas hypocarbia from hyperventilation results in vasoconstriction and a decrement in CBF with a subsequent reduction in ICP.¹⁶³ Even though pressure and metabolic autoregulation may be impaired after TBI, CO₂ reactivity often remains intact.^32,164

Though hyperventilation is an efficacious means of decreasing ICP, it does not have a role as a standard therapy in the management of patients with intracranial hypertension due to the potential of causing further brain injury secondary to ischemia. As noted earlier, studies using xenon CT^39,165 and implantable thermodiffusion blood flow probes¹⁶⁶ to measure CBF have demonstrated that patients with severe TBI have significant reductions in CBF in the first 24 hours after injury. Combined with evidence of cerebral ischemia seen in post mortem histologic analysis of patients with severe TBI, these data caution against aggressive hyperventilation in this patient population.^25,167,168 In turn, worse clinical outcomes are seen among severely head-injured patients who are subjected to prophylactic hyperventilation during the first 5 days after the injury.¹⁶⁹ Hence, hyperventilation should be avoided during the critical initial 24-hour postinjury period, and limiting the use of hyperventilation in patients with severe brain injuries could prevent iatrogenic cerebral ischemia from occurring and may potentially help improve neurologic recovery.¹⁷⁰ Though the routine use of hyperventilation is not indicated in the management of intracranial hypertension in patients with severe TBI, it may be used as a short-term maneuver to reduce ICP in the event of acute neurologic deterioration and impending herniation while other more definitive measures are taken.

Barbiturate-Induced Burst Suppression

The ability of barbiturates to lower ICP has been known since the 1930s.¹⁷¹ The mechanism by which this reduction in ICP occurs is not entirely understood but likely involves mediating cerebral vasoconstriction,¹⁷² reversibly reducing CMRO₂,¹⁷³ and preventing the onset of numerous excitatory neurotoxic cascades. Studies investigating clinical outcomes in patients who received barbiturate therapy have been inconclusive. Though prophylactic use of barbiturates has no clinical benefit,^135,174 Eisenberg et al.¹⁰⁶ illustrated that in the setting of intractable intracranial hypertension, patients whose ICPs were responsive to barbiturate therapy had a significantly reduced mortality.

The administration of barbiturates such as pentobarbital is generally performed concomitant with continuous electroencephalogram monitoring. Maximal reductions in CBF and cerebral metabolism are seen when electrographic burst suppression of cerebral activity is achieved.¹⁷⁵

A key side effect to consider prior to barbiturate administration is the potential to precipitate systemic hypotension. A meta-analysis by the Cochrane group noted that barbiturate infusion was associated with a fall in blood pressure in 25% of patients. In turn, the significant risk of causing hypotension—a factor known to be a harbinger of poor neurologic outcomes in patients with severe TBI—would offset any beneficial effect seen with ICP reduction.¹⁷⁶ No clear improvement in outcome has been seen in studies involving the use of barbiturate therapy, which has led to the Brain Trauma Foundation stipulating that high-dose barbiturate therapy should only be instituted in cases of treating intracranial hypertension refractory to all other medical and surgical treatment alternatives.¹²³ Furthermore, proper monitoring of cardiovascular status with a Swan-Ganz catheter and maintenance of hemodynamic stability are paramount if barbiturates are to be used.

Decompressive Craniectomy

Adjunctive Therapy: Lumbar CSF Drainage

As previously mentioned, cannulation of the ventricular system with an EVD allows for ICP monitoring and is a first-line therapy utilized to lower ICPs in patients with severe TBI, because it provides a mechanism by which ICP reduction is achieved through CSF diversion. Lumbar CSF drainage as a means of treating intracranial hypertension was heretofore considered dangerous, as initially noted by Cushing, because of its risk of precipitating transtentorial or tonsillar herniation.¹⁸⁹ Recent studies have posited that in cases of intracranial hypertension refractory to standard maneuvers directed toward lowering ICP, CSF drainage via placement of a lumbar drain is an effective means of achieving lasting ICP reduction, with a potentially acceptable risk of lethal outcome because of herniation at a rate of 6%. Improvement in neurophysiologic parameters such as P_btO₂ and CBF, as well as a lower mortality rate compared to historic cohorts among TBI patients with malignant ICPs treated with lumbar CSF drainage, has been seen.¹⁹⁰

Electrolytes and Fluid Balance

Upward of 60% of patients have electrolyte abnormalities following TBI.²⁰² Derangements in serum sodium levels are the most common and most pertinent from the standpoint of their role in maintaining neurologic homeostasis. Post-traumatic hyponatremia is associated with worse outcomes and longer hospital stays;²⁰³ thus, recognizing its etiology is crucial so that proper therapy may be directed accordingly. Hyponatremia commonly occurs in TBI patients as a consequence of the syndrome of inappropriate antidiuretic hormone (SIADH) or due to cerebral salt wasting. Management of SIADH includes fluid restriction and possible administration of demeclocycline. Patients with cerebral salt wasting often require sodium supplementation in the form of hypertonic saline or oral salt tablets. A key feature that distinguishes between these two entities is the patient’s fluid status: individuals with SIADH are euvolemic or hypervolemic, while those with cerebral salt wasting are generally hypovolemic. Regardless of etiology, slow correction of acute hyponatremia is a must to avoid iatrogenic central pontine myelinolysis.

Hypernatremia is also seen among patients who have suffered TBI. Central diabetes insipidus from pituitary stalk dysfunction often occurs with increased severity of brain injury and has been estimated to be the cause of hypernatremia in nearly 20% of cases, with significant implications on mortality.²⁰⁴ Administration of hypotonic fluids, desmopressin (DDAVP), and free water boluses may be considered in managing this problem.

Normal saline (0.9% NaCl) remains the standard choice of maintenance fluid in patients with head injury after initial resuscitation. As a rule, dextrose is excluded from the solution due to concern for causing hyperglycemia, which has been shown to worsen prognosis in patients with TBI.^13,205 Hypotonic fluids are not an acceptable choice as a maintenance fluid because of their potential to exacerbate cerebral edema. Fluid choice, however, must be dictated by the clinical situation, as solutions of varying tonicity have to be used in cases of SIADH, cerebral salt wasting, and diabetes insipidus.

In our institution, fluid and electrolyte management in patients with severe TBI is predicated on proper monitoring of cardiovascular parameters with central venous lines and indwelling arterial catheters. Thus, by tracking central venous pressures, we can gauge fluid status. Also, frequent checks of serum sodium concentration in the first 48 hours after the injury allow us to observe any trend toward an aberrant sodium level.

Glycemic Control

Elevated glucose levels are often seen among critically ill patients, regardless of whether they have a previous history of diabetes mellitus. Admission hyperglycemia has been demonstrated to correlate with increased risk of infection, longer lengths of stay in the intensive care unit and in the hospital, and higher mortality rates.^206–209 Tight glycemic control and maintenance of blood glucose below 110 mg/dl has been shown to decrease morbidity and mortality in intensive care patients; thus, intensive insulin therapy is a fundamental principle in the management of the critically ill.²¹⁰ Patients with severe head injuries commonly have supranormal glucose levels, likely secondary to a systemic surge in the sympathoadrenal axis and the resultant exaggerated catecholamine release.²¹¹ Hyperglycemia, both early and persistent, in patients with severe TBI has been shown to be associated with injury severity and portends a poor outcome.^13,205 The potential pathophysiologic process involves the development of intracellular acidosis secondary to the conversion of glucose into lactate via anaerobic metabolism.²⁰⁵ Frequent glucose monitoring and aggressive glycemic control, starting with administration of sliding scale insulin, to maintain blood glucose levels below 130 mg/dl are standard parts of the severe TBI protocol at our institution.

Hyperpyrexia

Fever increases cerebral metabolism. Secondary ischemia may occur when metabolic autoregulation is uncoupled and CBF does not match CMRO₂. Fever has been shown to be a significant predictor of increased mortality in patients with head injury⁶⁰ and must be aggressively controlled with antipyretics and cooling blankets. Aggressively pursuing an infectious workup to determine the etiology of fever is of the utmost utility.

Altered Endocrine Function

As mentioned previously, homeostatic mechanisms regulating water balance and electrolyte concentrations are frequently disturbed following TBI. Disorders of endocrine function secondary to damage to the anterior pituitary gland or the pituitary stalk result in aberrant circulating levels of hormones regulated by the anterior pituitary. Originally described in 1918,²¹² post-traumatic hypopituitarism was initially believed to be rare,²¹³ but more recent studies have shown widespread neuroendocrine dysfunction in the acute post-TBI period.^214,215 At our institution, a neuroendocrine panel is drawn on post-trauma days 3 and 7 in our severe TBI patients. Serum cortisol, insulin-like growth factor-1, free T4, thyroid stimulating hormone, estradiol and prolactin levels in females, and testosterone levels in males are checked. Of particular importance is adrenal insufficiency in the acute phase of TBI because of its potential to have life-threatening consequences in critically ill patients, manifesting as hyponatremia, hypotension, and hypoglycemia. Dramatic responses have been reported subsequent to glucocorticoid replacement in these individuals.²¹⁶ Early post-traumatic pituitary hypopituitarism resolves in a significant proportion of patients, generally by 6 months after the injury.²¹⁷

Post-traumatic hypopituitarism occurs in roughly 25% of long-term survivors of TBI,²¹⁷ with growth hormone and gonadotropin secretion being most commonly affected. Anterior pituitary hormone dysfunction measured more than 12 months after the injury has been reported to be as high as 59% among patients with severe TBI.²¹⁸ Other studies have noted that severe TBI, presence of higher ICP values, and longer intubation times are key factors associated with long-term prevalence of post-traumatic hypopituitarism.²¹⁹

Interestingly, post-traumatic pituitary dysfunction can evolve: patients who are not affected in the acute, post-injury period can exhibit hormonal deficiencies later.²¹⁷ Particular imaging findings, including CT evidence during the first 3 months after injury confirming the presence of brain swelling or a fracture of the cranial base or a MRI showing an empty sella within 12 months after the injury, are predictors of developing post-traumatic hypopituitarism.²²⁰

The presence of a neuroendocrine disorder is of tremendous clinical significance, not only from a general physiologic standpoint but also from a neurocognitive and neuropsychiatric standpoint.^221,222 In the presence of a concomitant neurologic deficit, the deleterious effect of a dysfunctional hypothalamic–pituitary axis on a patient’s quality of life and capacity for rehabilitation confers significant morbidity. Improved cognition and mental well-being seen in patients with post-traumatic hypopituitarism after hormone replacement reinforce the importance of diagnosis and treatment of this disorder. One such algorithm is seen in Fig. 132-11.

FIGURE 132-11 Algorithm for evaluating neuroendocrine function of patients following TBI.

(From Behan LA, Phillips J, Thompson CJ, et al. Neuroendocrine disorders after traumatic brain injury. J Neurol Neurosurg Psychiatry 79:753-759, 2008.)

Nutrition

Numerous studies over the past 25 years have highlighted the systemic hypermetabolic state induced in response to polytrauma and severe TBI. The overall state of catabolism is evidenced by patients’ supranormal caloric requirements in the acute, postinjury period: patients with severe head injuries expend upward of 140% to 160% of their basal rate, while those in barbiturate-induced burst suppression and patients who have undergone neuromuscular paralysis may require between 100% and 120% of normal resting levels.²²³

Nutritional support should be established as soon as possible, and patients should attain full caloric replacement by postinjury day 7.²²⁴ Severe undernutrition in the first 2 weeks after injury is associated with a much higher mortality rate as compared to the rate for patients who have had full caloric replacement by postinjury day 7.²²⁵ Protocols involving early initiation of feeding are correlated with decreased rates of infection and complications;²²⁶ regardless, to reach goal caloric replacement with the proper time afforded for advancement of parenteral or enteral nutrition, feeds should be started no later than 72 hours after injury.

The TBI literature is sparse with regards to research delineating the specific composition of feeds. To achieve an appropriate nitrogen balance, 15% of caloric intake should come from protein; otherwise, the appropriate amount of fats and carbohydrates constituting feeds has not been established among patients with severe brain injuries. The administration of glutamine, arginine, omega-3 fatty acids, nucleotides, and branch chain amino acids has shown promise in enhancing the immune response among critically ill patients,^227–229 but the applicability of the results specifically in the TBI population is unknown.

Gastric, jejunal, and parenteral methods of feeding represent the three options available for nutritional support. Orogastric tubes are preferred to the nasogastric route in the presence of skull base fractures or sinus disease. Altered function of the lower esophageal sphincter and slowed gastric emptying are mitigating factors in gastric feeding, especially when weighing the risk of aspiration. Long-term feeding may require placement of a percutaneous endoscopic gastrostomy tube or jejunal tube. Although the enteral route decreases the risk of bacterial translocation and sepsis, parenteral nutrition may serve as a temporary route for nutritional support until gastroparesis resolves.

Infections

Infections are important sources of morbidity and mortality in patients with severe TBI and, from a systems-based perspective, contribute significantly to healthcare costs, increasing patients’ length of stay in the hospital and requiring antibiotic treatment. Infectious considerations of special pertinence in the TBI population involve nosocomially acquired infections, as well as those associated with invasive cerebral monitoring.

Higher rates of systemic infection are correlated with increased severity of brain trauma, as well as the presence of concomitant chest trauma.²³⁰ Upward of 70% of mechanically ventilated patients can develop pneumonia.²³¹ The current recommendation of the Brain Trauma Foundation stipulates that administration of periprocedural antibiotics for intubation is warranted,²³² based on evidence showing a decrement in the incidence of pneumonia with such a protocol.²³¹ Unfortunately, prophylactic antibiosis during intubation in patients with severe TBI has not been shown to reduce mortality. Studies in which tracheostomy was performed within a week of injury similarly found no effect in reducing patient mortality in the severe TBI population, with no change in the rate of nosocomial pneumonia. The utility of early tracheostomy seems to lie in its reduction of the duration of mechanical ventilation in these critically ill patients.²³³ Caution is also warranted in considering prolonged antibiotic therapy in ventilator-dependent patients with severe TBI because of the potential to select for highly resistant bacterial species.

Invasive cerebral ICP monitoring confers a significant infection risk to the critically ill patient. Some studies have quoted rates of infection associated with placement of ICP monitoring devices as high as 27%.²³⁴ Interestingly, bacterial colonization of intraparenchymal monitors occurs at a higher rate than of EVDs.²³⁵ As previously mentioned, ventricular catheter exchange has not been found to be an efficacious method of reducing infection rates; in addition, prophylactic antibiotic use during ventricular catheter placement has not been recommended as a method of reducing infection. The key to avoiding infectious complications is predicated on using proper sterile technique when placing EVDs and other invasive intracerebral monitoring devices and, once the devices are placed, minimizing the frequency with which they are manipulated or flushed.

Pharmacologic Prophylaxis

Providing appropriate medical care to patients with severe TBI includes administration of pharmacologic agents to prevent potential complications. Gastric ulcers often form in patients after severe brain trauma, and in general, stress-induced ulceration of the gastric mucosa is a common complication among critically ill patients, with mortality rates of upward of 50% in patients requiring neurosurgical intensive care.²³⁶ Histamine type 2 (H2) antagonists or proton pump inhibitors are the most commonly used medicines for gastrointestinal prophylaxis, with recent data showing that both medications have a similar efficacy in preventing bleeding in the upper gastrointestinal tract in patients receiving mechanical ventilation.²³⁷ Given the relative cost effectiveness of H2 antagonists and the long-established effectiveness of these agents in the severe TBI population,²³⁸ the policy at our institution is to administer an H2 antagonist for gastric ulcer prophylaxis. Importantly, neutralizing gastric acid pH potentially places the patient at increased risk of developing nosocomial pneumonia; sucralfate, a medicine that strengthens gastric mucosa, may mitigate this risk.

The administration of anticonvulsants to prevent the development of post-traumatic seizures (PTSs) has been a mainstay in the management of brain trauma for the past several decades because, when not given pharmacologic prophylaxis, the incidence of clinically observed PTSs in the TBI population is relatively high: early PTS (those occurring within 7 days of injury) is estimated to occur in between 4% and 25% of patients, and late PTS (those occurring after postinjury day 7) happens in 9% to 42% of patients.^240–242 Risk factors for PTS development include the presence of cortical contusions; depressed skull fracture; subdural, epidural, or intracranial hematoma; penetrating injury; seizure within 24 hours of injury; or initial GCS score of fewer than 10 points.^242,243 Seizures increase ICP and CMRO₂; alter blood pressure, CBF, and oxygen delivery; and cause excess neurotransmitter release. Temkin et al.²⁴⁴ demonstrated that administration of phenytoin to brain-injured patients resulted in a significant decrement in the incidence of early PTS but not late PTS. Carbamazepine²⁴⁰ and valproic acid²⁴⁵ were found to have similar effects on the development of PTSs. The current recommendation is that anticonvulsants be administered for 1 week after the injury due to the potential for untoward side effects of treatment, including neurobehavioral consequences, and because antiseizure prophylaxis has not been shown to achieve a reduction in the incidence of late PTS.²⁴⁶ If late seizures occur, they are to be managed identical to the manner in which a first-time seizure in any other patient population is addressed. A recently published study comparing phenytoin and levetiracetam demonstrated that the latter agent was similarly effective in the prevention of seizures among the study population, 90% of whom were severe TBI patients, and was associated with improved outcomes and a significantly better side-effect profile.²⁴⁷ Thus, studies involving novel therapeutics and use of electroencephalography to detect subclinical seizures may provide further guidance in our management of PTS.

Prevention of Venous Thromboembolism

Patients with severe head injuries constitute a high-risk group for developing venous thromboembolic events (VTEs).²⁴⁸ In the absence of prophylaxis, studies have estimated the risk of deep venous thrombosis (DVT) among severe TBI patients to be upward of 20%.²⁴⁹ A distinction must be drawn between clinically symptomatic DVTs and those detected by screening tests, as DVTs detected through screening usually are relegated to the calf and are not associated with adverse consequences to the patient.²⁵⁰ DVTs involving the proximal leg veins, on the other hand, place the patient at increased risk for developing pulmonary embolism, which has been estimated to occur in 0.38% of TBI patients during their acute hospital stay.²⁵¹ The development of pulmonary embolisms has significant consequences on morbidity and mortality and important implications on the management of patients with TBI: the physician must weigh the safety of treating the thromboembolism with anticoagulant agents relative to the risk of progression of intracranial hemorrhage or may have to consider timing of anticoagulation as part of the patient’s postoperative management following craniotomy. Hence, prevention of VTE is of the utmost importance.

DVT prophylaxis can be accomplished via mechanical therapeutic modalities and by administration of pharmacologic agents. Mechanical therapies using compression stockings and intermittent pneumatic compression devices are efficacious in preventing DVTs.^252–254 They are associated with lowered risk of bleeding-related complications as compared to pharmacologic agents, but confounding factors among TBI patients such as multisystem trauma, including lower extremity injuries, may limit the applicability of this prophylactic measure to all patients with severe TBI. Administration of low-dose heparin or low-molecular-weight heparin is an effective means of preventing VTE²⁵⁵ and has a higher efficacy than use of mechanical modalities alone.^256,257 Pharmacologic prophylaxis, however, carries an increased risk of intracranial bleeding and should not be started in the perioperative setting. Though VTE prophylaxis was initiated in patients within 24 to 48 hours of admission in a number of studies, the exact time needed to establish a “safety window” following patients’ sustaining a TBI or after undergoing surgical intervention has not been defined. The policy at our institution to start upon admission VTE prophylaxis with intermittent pneumatic compression devices in all patients with severe TBI without extremity injuries precluding such an intervention, and pharmacologic prophylaxis is initiated 48 hours after either a CT scan demonstrates stable imaging findings or an operative intervention occurs.

Placement of a retrievable inferior vena cava (IVC) filter is another method to consider in preventing pulmonary embolism among trauma patients. Use of retrievable IVC filters also putatively spares patients from the untoward risks of anticoagulation associated with DVT treatment. Due to evidence of extremely low rates of filter retrieval, which are primarily attributable to poor patient follow-up but can involve technical issues, as well as development of significant thrombus within the IVC filter,²⁵⁸ we do not typically employ this modality in the management of patients with severe TBI.