Current Diagnosis & Treatment: Surgery

36

Neurosurgery

Aditya S. Pandey, MD
B. Gregory Thompson, MD

DIAGNOSIS & MANAGEMENT OF ALTERATIONS IN CONSCIOUSNESS

Kyle Sheehan, MD; George A. Mashour, MD, PhD

Alterations of consciousness and the related conditions of delirium, acute confusional state, and acute encephalopathy are among the most common mental disorders encountered in either surgical or medical patients. The prevalence of altered mental status in hospitalized patients is high, with reported rates up to 50%. These conditions are associated with increased mortality rates ranging from 10% to 65%, and excess annual health care expenditures in the billions. Given the high incidence of altered mental states, at least a basic understanding of the pathophysiology, diagnosis, and management of common etiologies is warranted for all medical practitioners.

DEFINITIONS

Consciousness is generally defined as the subjective experience of the environment and the self. It is comprised of two components: arousal, which is the state of wakefulness, and awareness, which is the state of phenomenal perception. This distinction is useful, since the two processes are dissociable. For example, a vegetative state is characterized by a patient that is awake (ie, the cortex is aroused), but not necessarily aware.

Arousal is generated by activity of the ascending reticular activating system, which is composed of neurons within the central mesencephalic brainstem, the lateral hypothalamus, and portions of the thalamus. Widespread projections of these nuclei synapse on neurons in the cerebral cortex and generate an arousal response. Arousal responses define the level of consciousness (eg, being awake vs. asleep vs. comatose). Awareness is thought to be generated through networks involving the thalamus and association cortices of the frontal, parietal, temporal, and occipital lobes. Processes related to awareness define the content of consciousness (eg, seeing a blue circle vs. a red triangle).

Many terms are used to describe the levels of consciousness ranging from alert to comatose. The alert patient is awake and immediately responsive to all stimuli. Stupor is a condition in which the patient is less alert but still responds with stimulation. An obtunded patient appears to be asleep much of the time but still responds to noxious stimuli. A vegetative state is a state of arousal without awareness in which the patient may open his or her eyes, track objects, chew, and swallow, but not respond to auditory stimuli or appear to sense pain (although pain processing is now known to occur in the vegetative state). The comatose patient appears asleep and does not respond to stimuli. Often, terms used to describe states of consciousness lack consistent definitions, and a clear description of a patient’s state of arousal and awareness results in more precise communication.

PATHOPHYSIOLOGY OF ALTERED CONSCIOUSNESS

In general, altered states of consciousness can arise from physiologic, pharmacologic, and pathologic causes. Before addressing pathologic causes (which can have structural or nonstructural etiologies), physiologic perturbations such as hypoglycemia, hypoxia, hypercarbia, hyponatremia, and hypothermia should be addressed. Pharmacologic etiologies, such as acute intoxication, overdose, and residual anesthesia from surgery, should also be considered and reversed when possible.

Structural Pathology

The reticular activating system is excited by a wide variety of stimuli, particularly somatosensory stimuli. Given that its nuclei are highly concentrated in the midbrain, it can be damaged by central midbrain lesions, which can result in the loss of arousal and coma.

Less severe dysfunction of the reticular activating system results in an acute confusional state. The cardinal signs of an acute confusional state are somnolence, inattention, and disorientation. Additionally, perceptions may be distorted, leading to hallucinations, and the patient may be unable to organize and interpret a complex stimuli. Disordered perception results in dysfunctional learning, memory, and problem solving. Thought processes can be disorganized and tangential, and the confused patient may develop delusions. In some cases, the acute confusional state presents as delirium, which can, in its hyperactive form, be characterized by heightened arousal, disordered perception, agitation, delusions, hallucinations, and autonomic hyperactivity (diaphoresis, tachycardia, hypertension, and mydriasis). However, it is important to note that hypoactive delirium is the most common type to be manifested in the postoperative setting.

Neurons in the pons, midbrain, and hypothalamus are necessary for regulation of sleep-wake cycles. Therefore, lesions involving the pons may preserve consciousness but disturb sleep. This is in contrast to the typical vegetative state that results from diffuse destruction or injury of the bilateral cerebral cortices, secondary to global cerebral ischemia or anoxia with preservation of the reticular activating system and brainstem sleep centers. This leaves the patient with preserved sleep-wake cycles without the ability to interact with the environment.

Acute confusional states and coma may result from structural or metabolic causes. Structural lesions of the cerebral hemispheres such as hemorrhage (intracerebral, subdural, or epidural), large areas of ischemic infarction, abscesses, or neoplasms can expand over minutes or a few hours and result in significant elevation of the intracranial pressure (ICP).

Nonstructural Pathology

Several nonstructural disorders that diffusely disturb brain function can produce a confusional state or, if severe, coma (Table 36–1). Acute toxic-metabolic encephalopathy (TME), which encompasses delirium and the acute confusional state, is an acute condition of global cerebral dysfunction in the absence of primary structural brain disease. Typically, TME is usually a consequence of systemic disease or secondary to drugs and metabolic toxins that are reversible, thus making prompt recognition and treatment critical.

Table 36-1. Etiology of altered state of consciousness.

Additionally, TME is common among critically ill patients and is probably underdiagnosed, especially when it occurs in patients who require mechanical ventilation. Risk factors for the development of TME include admission to an ICU, advanced age, preexisting primary neurodegenerative disease (dementia), nutritional deficiency, infection, temperature dysregulation, and failure of multiple organ systems.

PATHOPHYSIOLOGY OF ELEVATED INTRACRANIAL PRESSURE

The skull encloses three major components: brain parenchyma, cerebrospinal fluid (CSF), and blood. The volumetric sum of these components is maintained at a constant. Given the nondistensible cranial vault and these three noncompressible substances, an increase in one component must be offset by a decrease in another or pressure will increase. This offset is normally accomplished by displacement of CSF into the contiguous subarachnoid space of the spinal cord, termed “spatial compensation.” This delicate balance, however, has physiologic limits. Once these compensatory limits are met, an increase in the volume of any component will lead to an exponential increase in ICP as illustrated by the intracranial elastance curve (Figure 36–1).

Figure 36–1. Intracranial compliance curve demonstrating the approximate relationship between ICP and an expanding intracranial mass.

Normal ICP in adults is less than 10-15 mm Hg. The presence of tumor or blood, the impedance of CSF drainage (hydrocephalus), cerebral edema, and increased cerebral blood flow (hyperemia as a response to head injury or hypoventilation causing hypercarbia and vasodilation) may raise the ICP to a dangerous level. The increased ICP may lead to reduced cerebral perfusion and ischemia and a mass effect progressing to shifting of the brain parenchyma and devastating herniation.

SIGNS & SYMPTOMS OF INCREASED ICP

Increased ICP is often accompanied by altered mental status, headache, vomiting (without nausea), and papilledema on fundoscopic examination. Cushings triad may be seen in patients with severely increased ICP and consists of hypertension, bradycardia, and respiratory irregularity. Obtundation, focal neurologic findings including unilateral pupillary dilation from pressure or traction exerted on the third cranial nerve (ie, the “blown” pupil), and hemodynamic instability are late findings and indicate pending uncal herniation.

Cerebral Herniation Syndromes

Cerebral herniation occurs when increased ICP results in a portion of the brain shifting from one intracranial compartment to another. These shifts may lead to vascular compromise and infarction of portions of the brain, disruption of white matter pathways, and direct pressure on structures such as cranial nerve III.

Subfalcine Herniation

Occurs in patients with a frontal lobe mass as the cingulate gyrus herniates beneath the falx. Symptoms are often related to the mass or increased ICP.

Lateral Mass Herniation

Occurs in patients with an expanding lateral mass. Symptoms include contralateral hemiparesis, diminished consciousness, and an ipsilateral cranial nerve III palsy. Compression of pupillary fibers also causes a dilation of the pupil and as lateral displacement of the midbrain continues there is ipsilateral hemiplegia. Late in the process, the uncus and hippocampus may herniate transtentorally.

Cerebellar Tonsillar Herniation

Posterior fossa masses cause symptoms by compression of the brainstem and obstructing the flow of CSF (hydrocephalus). As pressure increases, the cerebellar tonsils may be pushed into or through the foramen magnum. As the medulla is compressed, apnea results from dysfunction of the medullary respiratory center.

Upward Transtentorial Herniation

Posterior fossa masses may cause obstructive hydrocephalus. If these patients undergo ventriculostomy, there is a possibility of upward herniation of posterior fossa contents into the region of the thalamus and hypothalamus.

DIAGNOSTIC EVALUATION OF ALTERED CONSCIOUSNESS

The diagnosis of an altered state of consciousness is provided by general physical examination, neuroradiological imaging, drug screens, and certain laboratory studies.

Neurological Examination

The neurologic exam includes assessment of level of consciousness, brainstem reflexes, and motor activity. Rapid, standardized assessment of the level of consciousness is useful for clinical decision-making and communication. The level of alertness reflects the severity of the underlying condition; severely affected patients are comatose. Commonly, the Glasgow Coma Scale (GCS) (Table 36–2) is used to assess level of consciousness in patients with head injury, but is also widely used as assessment of level consciousness regardless of etiology. A patient with a GCS score 8 or less is generally considered to be comatose. The cardinal feature of confusion and delirium is impaired attention. Simple bedside tasks such as serial subtraction or naming the months of the year in reverse can test attention. Marked fluctuations in mental status over time are characteristic. Other common findings include a disturbed sleep-wake cycle, decreased alertness, hypervigilance, hallucinations, sensory misperceptions, impaired memory, and disorientation. The thought process is often disorganized, manifested by confused or rambling conversation.

Table 36-2. Glasgow coma scale.

The cranial nerves and their corresponding brainstem reflexes, including pupillary response to light, corneal reflexes, oculocephalic reflex (doll’s eyes), vestibulo-ocular reflex (cold calorics), breathing patterns, cough, and gag reflexes are further used to describe accurately the neurologic status and localize the level of brainstem dysfunction. Generally, the brainstem reflexes are only affected in severe TME. Abnormal patterns of respiration, particularly Cheyne–Stokes respiration, may also occur with significant brainstem dysfunction.

Motor responses may be delineated as spontaneous or induced by noxious stimuli, purposeful or nonpurposeful, unilateral or bilateral, and upper or lower extremity. The patient may display withdrawal from a stimulus, abnormal flexion (decorticate posturing), abnormal extension (decerebrate posturing), or absence of motor activity. A variety of motor abnormalities are also associated with TME: tremor, asterixis, multifocal myoclonus (sudden, nonrhythmic gross muscle twitching), paratonia (increased tone with variable resistance), diffuse brisk deep tendon reflexes, or bilateral extensor plantar responses.

Laboratory Studies

The laboratory investigation of decreased level of consciousness includes a complete blood count, coagulation studies, electrolyte panel, and examination of calcium, magnesium, phosphate, glucose, blood urea nitrogen, creatinine, bilirubin, liver enzymes, ammonia, serum osmolality, and arterial blood gases. Toxicology screening should be performed for suspected intoxications. Blood and CSF cultures should be obtained if infection appears present. Further analysis of CSF with cell counts, protein, glucose, and organism-specific studies such as herpes simplex virus PCR and fungal serology can reveal meningitis, neoplastic cells, or subarachnoid hemorrhage. Thyroid function tests and vitamin B12 and serum cortisol concentrations should be assessed if endocrinopathy is considered.

Neuroradiological Studies

Emergent computed tomography (CT) imaging to assess for structural lesions (intracerebral hemorrhage, large territory ischemic stroke, hydrocephalus, neoplasm, and diffuse cerebral edema) is strongly recommended for patients presenting with focal neurological deficits or severe decline in the level of consciousness. Acute ischemic stroke within the first 3-4 hours may be clinically occult on CT imaging. Magnetic resonance imaging (MRI) of the head is generally reserved to further assess for acute ischemic stroke as well as further characterization of neoplastic lesions. MRI should be not relied upon for an emergent diagnosis given the duration of the study.

Electroencephalography

The electroencephalogram (EEG) can both confirm global cerebral dysfunction and exclude subclinical status epilepticus with greater sensitivity than clinical examination alone. After structural lesions have been excluded, EEG should be performed in most patients with altered consciousness at some point during the course of their care. The degree of diffuse slowing of the normal background plus abnormal mixed rhythms in the EEG correlates with the severity of TME. Slowing can be categorized as follows: mild, with a reduction in the normal alpha frequencies (8-13 Hz); moderate, with theta frequencies (4-8 Hz); severe, delta frequencies (less than 4 Hz).

ICP Monitoring

Quantification of ICP is critical for management of patients with large intracerebral lesions resulting in mass effect and shift of midline structures, diffuse cerebral edema, or hydrocephalus. ICP cannot accurately be estimated based on clinical findings or imaging. There are several methods of direct ICP measurement, but the two most commonly employed in clinical practice are ventricular catheters and intraparenchymal microtransducer systems. Other methods such as subarachnoid and epidural devices have much lower accuracy.

The gold standard of ICP monitoring is via an intraventricular catheter connected to a standard pressure transducer. These catheters are usually placed into the lateral ventricle by a small frontal burr hole. The advantages of intraventricular catheters are that they measure global ICP, allow for therapeutic drainage of CSF, and are amenable to external calibration. Disadvantages are risk for infection, hematoma, or difficult insertion.

Microtransducer-tipped ICP monitors are placed in the brain parenchyma or subdural space either through a skull bolt, a burr hole or intraoperatively. They are almost as accurate as intraventricular catheters and have the advantages of lower infection and complication rates. The major disadvantages are an inability to drain CSF, no in vivo calibration and a small zero drift over time.

MANAGEMENT OF ALTERED CONSCIOUSNESS

The initial treatment of a patient with an alteration in consciousness must employ management of the ABCs: airway, breathing, and circulation. Patients who are not responsive enough to protect their own airway require intubation to reduce the risk of aspiration of gastric contents. During intubation, hypoxia and hypotension should be avoided in brain-injured patients and a high suspicion of cervical spine injury should be maintained. A patient presenting with a GCS score 8 or less should be intubated. Mechanical ventilation should be used if necessary to provide adequate oxygenation and ventilation as guided by arterial blood gases. Patients presenting with hypotension and shock must be aggressively treated with fluids and vasopressors to maintain adequate perfusion. Sources of shock (septic, cardiogenic, hypovolemic) should be investigated and treated appropriately.

Regardless of the cause of acute decline of mental status, a number of general measures should be instituted. There should be a discontinuation of all drugs with potential toxicity to the central nervous system, if possible. Antipsychotic medications such as haloperidol or quetiapine can be used for management of severe agitation. Thiamine should be administered to patients with a history of alcoholism, malnutrition, cancer, hyperemesis gravidarum, or renal failure on hemodialysis in order to prevent the development of Wernicke encephalopathy.

In the event that a patient presents with history or examination findings suggestive of elevated ICP, immediate life-saving measures may be required prior to a more detailed workup with neuroradiological imaging or ICP monitoring. These situations generally rely upon clinical judgment and expert consultation with a neurosurgeon or neurologist is strongly recommended. An examination consistent with a critically elevated ICP (coma with GCS score < 8, unilateral or bilaterally fixed and dilated pupil(s), decorticate or decerebrate posturing and Cushing triad of bradycardia, hypertension, and respiratory depression) warrants immediate intervention while delaying additional diagnostic studies. In addition to standard resuscitation measures, elevation of the head above the heart (usually 30 degrees) to increase intracranial venous outflow, temporary hyperventilation to a goal PaCO2 of 26-30 and administration of intravenous hyperosmolar therapy (either mannitol 1-1.5 g/kg or hypertonic saline infusion). Immediately following these measures, rapid evaluation of the underlying diagnosis by the patient history, detailed neurological examination and neuroradiological imaging should be pursued. Neurosurgical consultation should be considered for potential placement of a ventriculostomy as a means to assess ICP and potentially treat with CSF drainage.

If elevated ICP is present, therapy should be directed to maintain ICP less than 20 mm Hg. Interventions should be utilized only when ICP is elevated more than 20 mm Hg for more than 5 minutes. Transient physiologic elevations in ICP may be observed in the setting of coughing, movement, suctioning, or ventilator asynchrony that should not be targeted by therapy. During periods of elevated ICP, it is important to maintain adequate mean arterial pressure to ensure adequate cerebral perfusion pressure (which is mean arterial pressure – ICP).

GENERAL MANAGEMENT OF ELEVATED ICP

  1. Blood pressure control—Therapy with vasopressors should be targeted to maintain an adequate cerebral perfusion pressure (CPP = MAP − ICP), typically more than 60 mm Hg. Hypertension should generally only be treated when CPP more than 120 mm Hg. CPP more than 50 mm Hg is associated with cerebral ischemia.
  2. Position—Patients with elevated ICP should be positioned with the head elevated 30 degrees above the heart while maintaining the neck in neutral position without excessive flexion or rotation to maximize intracranial venous outflow.
  3. Removal of CSF—When hydrocephalus (either obstructive or communicating) is discovered, a ventriculostomy may be utilized to reduce intracranial CSF volume and secondarily ICP. Controlled drainage of CSF at a rate of approximately 1-2 mL/min, at intervals of a few minutes until the goal ICP is reached (ICP < 20 mm Hg) or until CSF is no longer easily obtained is recommended. CSF drainage via an intrathecal lumbar drain is generally contraindicated in the setting of elevated ICP due to the risk of transtentorial herniation.
  4. Hyperventilation—The use of mechanical ventilation to lower PaCO2to a goal of 26-30 mm Hg will induce cerebral vasoconstriction, decrease the cerebral blood volume, and rapidly reduce ICP. The effect of hyperventilation on ICP typically lasts for a period of hours prior to metabolic compensation. Therapeutic hyperventilation should be utilized only as an emergent intervention to temporarily control ICP and be replaced by other therapy modalities.

HYPEROSMOTIC THERAPY

  1. Mannitol—Osmotic diuretics reduce brain volume by creating an osmotic gradient from the brain parenchyma to the intravascular space, thus drawing free water from the parenchyma. Mannitol, prepared in a 20% solution, can be given as a bolus of 1-1.5 g/kg. Mannitol can be used in serial doses at intervals of every 6-8 hours at a reduced dosage on 0.25-0.5 g/kg as needed for continued elevated ICP. The onset of action is within minutes and the duration of effect varies widely from 4 to 24 hours. Serial measurements of serum sodium, serum osmolality, and renal function are necessary to prevent overdosage. Contraindications to the use of mannitol include serum sodium more than 150 mEq, serum osmolality more than 320 mOsm, or evidence of evolving acute tubular necrosis. In addition, mannitol frequently can induce hypotension and subsequently cerebral perfusion. Mannitol should not be used in patients with acute or chronic renal disease.
  2. Hypertonic saline—Bolus dosing of hypertonic saline (tonicity ranging 1.8-23.4 percent) can acutely lower ICP. The typical volume of hypertonic saline varies widely depending on tonicity, ranging from 30 mL of 23.4% to 1 L of 1.8%. Continuous infusion of hypertonic saline (1.8%-3%)to maintain hypernatremia may also be effective in controlling ICP. A meta-analysis of multiple clinical trials comparing the efficacy of mannitol to hypertonic saline for management of elevated ICP from a variety of causes (traumatic brain injury, stroke, tumors) found that hypertonic saline appeared to have greater efficacy in reducing elevated ICP, but clinical outcomes were not examined. Hypertonic saline should generally not be administered through a peripheral intravenous catheter.
  3. Fluid management—Typically, patients with elevated ICP should be kept euvolemic and normo- to hyperosmolar. Administration of free water and hypotonic solutions should be strictly avoided and instead isotonic fluids should be used in all maintenance fluids and infusions. Serum sodium levels should be closely monitored and hyponatremia should be appropriately corrected.
  4. Sedation—Maintenance of adequate sedation can decrease ICP by reducing cerebral metabolic demand, ventilator asynchrony, and the sympathetic responses of hypertension and tachycardia. Adequate sedation typically requires the establishment of a secure airway. Infusion of propofol is often the drug of choice for sedation because it can be rapidly titrated, thus allowing for frequent neurologic assessments.
  5. Fever—Fever increases brain metabolism, which increases cerebral blood flow and thus elevates ICP. Additionally, fever has been demonstrated to worsen brain injury in animal models. Therefore, aggressive treatment of fever, including acetaminophen and cooling, is recommended in patients with increased ICP.
  6. Antiepileptic therapy—Seizures, either convulsive or nonconvulsive, increase cerebral metabolism and result in ICP elevation. Aggressive treatment of seizures with antiepileptic therapy or infusion of anesthetics as well as continuous EEG monitoring is warranted. There is no clear evidence that prophylactic antiepileptic therapy is of any clinical benefit; however, consideration of prophylactic antiepileptic therapy is reasonable when high-risk mass lesions, such as those within supratentorial cortical locations, or lesions adjacent to the cortex, such as subdural hematomas or subarachnoid hemorrhage are present.
  7. Glucocorticoids—Glucocorticoids, typically dexamethasone, are reserved for the management of elevated ICP secondary to vasogenic edema secondary to intracranial neoplasm and infection. Typically, dexamethasone is dosed every 6-12 hours at a wide range of dosages. The use of glucocorticoids has been associated with a worse outcome in a large randomized clinical trial in traumatic brain injury and is no longer recommended. Glucocorticoids are not considered to be useful in the management of cerebral infarction or intracranial hemorrhage.
  8. Barbiturates—The use of barbiturates to control ICP is based on the drug’s ability to dramatically reduce brain metabolism and secondarily cerebral blood flow. Pentobarbital is most commonly used, with a loading dose of 5-20 mg/kg as a bolus, followed by 1-4 mg/kg/h. The barbiturate infusion is titrated based on assessment of ICP, CPP, and the tolerance of side effects. Continuous EEG monitoring is generally used, with titration to an EEG burst suppression pattern indicating appropriate suppression of cerebral metabolism. Barbiturate therapy is complicated and fraught with complications, particularly hypotension, adynamic ileus, reduced mucociliary clearance of the airway, and high risk of infections. In general, the use of barbiturates is reserved for elevated ICP refractory to all other treatment modalities.
  9. Therapeutic hypothermia—The use of hypothermia to treat elevated ICP has been controversial for decades, and its use is not recommended as a standard treatment for increased ICP. Hypothermia decreases cerebral metabolism and secondarily reduces cerebral blood volume and ICP. When used, hypothermia is achieved by whole body cooling using surface or intravascular cooling devices to a goal of 32-34°C. Studies have demonstrated significant side effects, including cardiac arrhythmias and severe coagulopathy. Given the multiple uncertainties of the appropriate use of therapeutic hypothermia in patients with elevated ICP, this treatment should be limited to patients with intracranial hypertension refractory to other therapies.
  10. Decompressive craniectomy—Decompressive craniectomy is the surgical removal a large portion of the cranial vault to allow for the edematous intracranial contents to expand and subsequently reduce ICP. At the time of the procedure any mass lesion (neoplasm or hematoma) is also removed. Decompressive craniectomy has been considered a last resort; some evidence suggests that it does improve outcomes.

Bhatia A, Gupta AK: Neuromonitoring in the intensive care unit. I. Intracranial pressure and cerebral blood flow monitoring. Intensive Care Med 2007;33:1263-1271.

Diedler J, Sykora M, Blatow M, et al: Decompressive surgery for severe brain edema. J Intensive Care Med 2009;24:168.

Edwards P, Arango M, Balica L, et al: Final results of MRC CRASH, a randomised placebo-controlled trial of intravenous corticosteroid in adults with head injury-outcomes at 6 months. Lancet2005;365:1957.

Kamel H, Navi BB, Nakagawa K, et al: Hypertonic saline versus mannitol for the treatment of elevated intracranial pressure: a meta-analysis of randomized clinical trials. Crit Care Med 2011;39:554.

The Brain Trauma Foundation; American Association of Neurological Surgeons; Congress of Neurological Surgeons; Joint Section on Neurotrauma and Critical Care: Guidelines for the management of severe traumatic brain injury. VI. Indications for intracranial pressure monitoring. J Neurotrauma 2007;24:S37-S44.

IMAGING OF THE CENTRAL NERVOUS SYSTEM

Douglas J. Quint Shawn L.A. Hervey-Jumper

INTRODUCTION

Imaging has become central to medical care in the past 25 years and is one of the fastest growing components of medical care expenditures in this country. In no field of medicine has imaging made more of an impact than in the neurosciences. Our ability to demonstrate anatomy and pathology noninvasively and to treat many pathologic central nervous system (CNS) processes utilizing minimally invasive techniques has grown immensely with improvements in imaging technology in the last few years.

Forty years ago, the only way to image the CNS was to replace cerebrospinal fluid (CSF) with contrast material (dye) or air via a lumbar puncture and take x-rays—a painful technique called pneumoencephalography—normal and abnormal CNS structures could be crudely outlined in this manner. Pneumoencephalography was abandoned in the late 1970s with the development of computed tomography (CT). Injecting dye directly into major blood vessels of the neck (cerebral angiography) has been available for three quarters of a century and allows exquisite delineation of intrinsic vascular pathology, but still does little to directly visualize the brain or spinal cord. While angiography is still performed today for both diagnostic and therapeutic purposes, the catheter is now placed via femoral arterial cannulation instead of directly into a neck blood vessel.

With the advent of CT scanning in the early 1970s, direct visualization of the brain was finally possible though resolving different normal brain structures and some pathologic processes remained difficult. Magnetic resonance imaging (MRI) first became available for clinical use in the early 1980s and has been the standard for evaluation of most CNS processes since the mid-1980s.

CT and MRI have continued to mature through the 1990s and into the 21st century. In addition to the introduction of lower radiation dose scanner capabilities on the most recent generation CT scanners, rapid scanning techniques on such scanners (“multislice” or “multidetector” CT scanners) now allow for collection of data reflecting cerebral blood perfusion, which can be collected during a 5-minute study. Similarly, such state-of-the-art scanners can be used to generate models of the major blood vessels of the brain (CT angiograms) lessening the need for more invasive intravascular catheterization in some patients.

MRI advances have been even more dramatic particularly with the release of higher field strength MR scanners (3.0 T) for clinical use in 2005. For example, the sensitivity of some of the newer MRI scanning techniques enables identification of physiologic changes in the brain minutes after ischemia occurs, a time frame that can potentially permit pharmacologic interventions that can affect clinical outcomes (eg, stroke). Functional MRI can assess eloquent areas of the brain near pathologic lesions (eg, tumors) that ideally should be avoided during surgery. MR spectroscopy (MRS) can identify lesion metabolites that can help discriminate among pathologic processes (eg, tumors vs. necrosis vs. ischemia vs. inflammation vs. infections). MRS can also identify metabolites accumulating in the brain in patients with congenital metabolic disorders. MRI can also be used to generate angiographic (MR angiography, MRA) images to evaluate blood vessels without injection of dye or the use of ionizing radiation.

Catheter (endovascular) angiography has been available in various forms for over 80 years, and has also continued to evolve over the last decade. Endovascular surgery—treating blood vessel abnormalities such as aneurysms or arteriovenous malformations (AVMs) through a catheter instead of with open surgery—is now commonly performed and has become the treatment of choice for the management of many lesions. More than half of aneurysms treated in the United States are currently managed endovascularly avoiding the need for craniotomy in the majority of those patients. Similarly, in the setting of an acute stroke due to an obstructing thromboembolus, in an attempt to reopen the involved vessel, a catheter can be placed in the region of the obstructing lesion and a clot mechanically removed through the catheter, or clot-dissolving (eg, thrombolytic) agents injected.

BASIC CENTRAL NERVOUS SYSTEM IMAGING TECHNIQUES

  1. Plain radiographs (“x-rays”) are relatively inexpensive, universally available and can demonstrate osseous abnormalities such as fractures or gross destructive lesions (Figure 36–2). The main disadvantage of plain x-rays is that essentially all normal soft tissues (eg, all intracranial and spinal canal structures) and pathologic processes (eg, hemorrhage, infarctions, tumors, abscesses, herniated disks, etc) cannot be detected. Even many intrinsic osseous lesions are difficult to delineate. In fact, until 30%-50% of bone marrow trabeculae is replaced by a pathologic process (eg, tumor, infection), no osseous abnormality is seen on a plain radiograph.

Figure 36–2. Lateral plain radiograph (x-ray) of the skull (ionizing radiation). Note the excellent delineation of the osseous structures, but poor delineation of soft tissues; specifically, no portion of the brain is imaged. Also, this study is not tomographic (ie, is not a slice) and therefore both the left and right sides of the head are superimposed.

  1. Another limitation of plain radiographs is that all structures are superimposed on an single image as the x-ray beam passes through the entire head or spine (as opposed to the “slices” generated on CT and MR scans).
  2. There is little role for plain radiographs for the evaluation of CNS disease as the interior of the head or spinal canal is not imaged. One exception is in patients with suspected child abuse. In these patients, in addition to other cross-sectional imaging (such as CT/MR which are performed to assess for intracranial injuries), plain radiographs shouldalso be obtained as they may show subtle nondisplaced fractures not readily identified on other imaging tests.
  3. Other roles for plain radiographs include assessing spinal motion (ie, lateral flexion/extension radiographs of the cervical or lumbosacral spine to assess spinal stability), localization of foreign bodies, and assessing intracranial or spinal structures where internal fixation hardware and/or other radiodense/ferromagnetic foreign material is present (which can limit the use of CT or MRI scanning).
  4. Ultrasound of the CNS, beyond the neonatal period, is predominantly used in the intraoperative setting to evaluate ventricular morphology and underlying parenchymal pathology. Intraoperative ultrasound can be used to help position ventricular catheters during placement of shunts. It can also be used to guide intracranial and intraspinal tumor resections.
  5. Outside of the operative setting, ultrasound has growing applications because it is noninvasive, portable, and does not involve radiation. However, an “acoustic window” is necessary for viewing the intracranial/intraspinal tissues meaning that overlying osseous structures usually must be removed before scanning can be performed (ie, portions of the skull or posterior spinal elements must be removed before the desired regions of the brain or spinal cord, respectively, can be evaluated).
  6. Ultrasound has multiple applications in infants (usually such infants still have open fontenelles so no additional “acoustic window” to the brain needs to be created). It has become the imaging study of choice for the evaluation of premature infants to assess for intraventricular hemorrhage and/or hydrocephalus. In adults, intracranial Doppler examinations can be performed by scanning through sutures. For example, scanning through the temporal suture allows access to the circle of Willis. These examinations can be performed at the bedside for assessment of arterial vasospasm (eg, in the setting of subacute subarachnoid hemorrhage). Ultrasound can also be used in younger children to identify the conus medullaris.
  7. CT scanning without or with intravenous administration of iodinated contrast material is performed utilizing a thin fan-like band of x-rays (ionizing radiation) that are generated by an x-ray tube that literally moves around the patient in approximately 1 second. The resultant image, which represents a “slice” of tissue, can be obtained with imaging thickness as thin as 1 mm. “Reformatted” submillimeter images can also be generated by the CT computer from the primary axially obtained CT scan data in any plane (sagittal, coronal, off-axis, etc). Three-dimensional images can also be created in this manner.
  8. CT demonstrates most osseous abnormalities better than any other imaging test with the possible exception of nondisplaced, nondistracted fractures which are still often better seen on plain x-rays (Figure 36–3). Soft tissue lesions can also be detected with much better sensitivity than plain x-rays. However, resolving some similar, but not identical, normal soft tissues from one another, delineating normal from pathologic soft tissue, and evaluating certain areas of the brain which are limited by scanning artifacts can still be difficult with CT scanning which is why CT remains inferior to MRI for evaluation of most brain and spinal canal abnormalities.

Figure 36–3. CT scan (ionizing radiation). This image represents a single section (“slice”) through the head (ie, it is a tomographic image; the left and right sides of the head can be separately delineated). The brain is directly imaged though resolving differences in intracranial structures can be difficult (eg, resolving gray and matter structures).

  1. When discussing CT images, the terms “density” and “attenuation” refer to the same process, namely, absorption of the x-ray beam. Areas of increased “density” have greater “attenuation” of the x-ray beam and result in more whitishareas on the CT scan image (eg, bones, iodine-based contrast material, acute blood, areas of calcification, and some foreign bodies). Areas of lower density have lower “attenuation” as they absorb less of the x-ray beam as it passes through the patient resulting in darker areas on the CT scan image (eg, fluid in the ventricles, gas, and fat).
  2. CT image contrast can be adjusted at a workstation after scans have been obtained to highlight differences in the densities of different tissues.
  3. Finally, the recent development of new methods of reconstructing CT data has resulted in significant reduction in patient radiation dosage exposure during many CT scans.
  4. Myelography involves performing a lumbar subarachnoid puncture (or a lateral C1-2 subarachnoid puncture) and instilling less than an ounce of iodinated contrast material to opacify the spinal canal subarachnoid space which results in outlining cauda equina nerve roots and the spinal cord, and any process that impinges upon (or is within) the subarachnoid space. Such processes include extrathecal tumors, infections, herniated disk material, degenerative spinal changes, and also processes that directly involve the intrathecal structures (spinal cord tumors, vascular malformations, metastases, etc).
  5. Myelography is no longer a primary imaging technique for evaluating the spinal canal having been replaced by MRI. It is currently reserved for “problem-solving” such as when the results of an MRI scan are not clear or when MR imaging is not possible (contraindications to MR scanning, internal fixation hardware limits evaluation by MRI, etc).
  6. Myelography is essentially always followed immediately (within hours) by CT scanning to better delineate relationships of pathologic processes to the subarachnoid space.
  7. MRI is an imaging technique that does not use ionizing radiation.
  8. The physics of creating an MR scan are quite complex. Briefly, a patient is placed in a strong magnetic field (30,000 × that of the earth’s magnetic field). Radiofrequency pulses are transiently (milliseconds) applied to the patient to briefly perturb the patient’s water molecules by raising them to a slightly higher energy level (quantum mechanical model). After turning off the radiofrequency pulse, these water molecules rapidly return to their respective baseline states by giving off their recently absorbed energy. The rate at which these perturbed water molecules return to their respective baseline states can be measured by using extremely sensitive receiver coils in the MR scanner. As the rates at which these molecules return to their respective baseline states varies by local magnetic environment (eg, the local magnetic environment is different in the ventricular system as opposed to the lentiform nucleus, white matter, the eyeball, muscles, tumors, etc), these detectable differences among tissues can be localized in 3-dimensional space and used to create an image.
  9. MR images are obtained which highlight different kinds of magnetic field differences between tissues. In general, most MR studies include “T1-weighted” and “T2-weighted” scans as part of the overall evaluation of the patient. MR scans take longer to perform than CT scans in part because T1-weighted (T1w) and T2-weighted (T2w) scans and often additional scans must be obtained separately. T1w images are distinguishable by the black appearance (“absence of MR signal”) of the CSF over the surface of the brain and in the ventricles. In contrast, on a T2w scan, the CSF over the surface of the brain and in the ventricles is white (Figure 36–4).

Figure 36–4. MR scan (no ionizing radiation). Similar to CT, this image represents a single section (“slice”) through the head. The white matter and gray matter structures are more easily resolved that on a CT scan, and even some gray matter structures can be resolved (eg, putamina and globus pallidae).

  1. In general, T1w scans are best for delineating anatomy and areas of contrast enhancement. T2w scans are exquisitely sensitive to subtle changes in water concentrations (in both normal and pathologic tissues). These changes are seen in most pathologic processes (eg, strokes, tumors, infections). These subtle changes are manifest as increased signal (more whiteness) on T2w scans.
  2. In the past 10 years, most MR studies include “FLAIR” (FLuid Attenuated Inversion Recovery) scans. In general, FLAIR scans can be considered super T2w scans on which normal fluid (eg, in the ventricles and subarachnoid spaces) has no signal (ie, black), and therefore, subtle pathologic process (which still appear white on the FLAIR scans similar to that seen on standard T2w scans) are easier to identify.
  3. A gadolinium-based contrast agent can be injected intravenously to enhance the appearance of some pathologic processes such as those outside the blood-brain-barrier (BBB) or are intraaxial but do not have a functioning BBB.
  4. MRI best delineates soft tissues both intrinsic and immediately extrinsic to the brain, can be performed in any plane (including nonorthogonal planes), and has no known side effects at the field strengths used for clinical studies. In addition to better demonstrating tissues than CT, MRI is not limited by many of the artifacts that limit evaluation by CT, particularly in the posterior fossa and spinal canal regions where CT can be severely limited. While not ideal for directly visualizing dense osseous structures (where many fractures, cortical erosions, and degenerative changes occur), MR is superb for detecting intrinsic (eg, bone marrow) osseous abnormalities (eg, metastatic disease, discitis/osteomyelitis, etc).
  5. Cerebral angiography is performed by directly placing a catheter into the femoral artery, passing it cephalad in a retrograde manner up the aorta to the level of the aortic arch, manipulating it into either a vertebral or carotid artery and then further cephalad into the neck and even intracranially (Figure 36–5). The catheter is moved utilizing intermittent fluoroscopic guidance with small amounts of dye injected at selected intervals to confirm the location of the catheter. When in position, larger amounts of dye are injected and serial plain digital x-rays are rapidly exposed as the dye passes through the intracranial vessels in the distribution of the injected blood vessel. Over 100 images might be obtained during a single 8-second injection of contrast material. This procedure is associated with a 0.1%-0.5% chance of causing a stroke.

Figure 36–5. Angiogram (ionizing radiation). A catheter was placed in the right carotid artery in the neck and dye injected while imaging was performed in a front-to-back projection of the right side of the head. The bones are then digitally “subtracted” from the image. The exquisite delineation of the blood vessels of the right carotid territory are demonstrated, but no intra-axial (ie, brain soft tissues) structures.

  1. This technique best delineates intrinsic blood vessel pathology such as atherosclerosis, aneurysms, AVMs, fistulas, vasculitis, blocked vessels, and other intrinsic vascular disorders. It can be used to inject medications into specific blood vessel territories for diagnostic or therapeutic purposes. Catheters can be placed directly into aneurysms, vascular malformations, and recently occluded vessels for definitive therapy.
  2. Endovascular (catheter) angiography is risky, expensive, requires complex equipment, and is performed by highly trained personnel. There are currently several alternatives to catheter angiography to evaluate for vascular lesions such as atherosclerotic vascular narrowing and aneurysms. Ultrasound (in the neck), CT angiography, and MR angiography are each useful alternatives in many situations (Figure 36–6).

Figure 36–6. Aneurysm demonstrated by multiple imaging techniques. A. Catheter angiogram (ionizing radiation). B. CT angiogram (ionizing radiation). C. MR angiogram (no ionizing radiation). Note that the catheter angiogram (A), an imaging modality that is associated with a risk of causing a stroke, best defines this anterior communicating artery region aneurysm (arrow). However, the aneurysm is still well-seen on the CT angiogram (B), an imaging technique that is not associated with any risk of causing a stroke, and is also well-seen on the MR angiogram (C), an imaging technique that is not associated with any risk of causing a stroke and does not employ ionizing radiation.

  1. Radionuclide Imaging (Scintigraphy) includes positron emission tomography (PET) and SPECT (singlephoton emission CT) scans which are molecular imaging techniques. Unlike other imaging techniques, these imaging modalities provide information beyond the structural appearance of normal and pathologic tissues. These techniques can provide physiologic information reflecting the functional state of tissues.
  2. Low doses of radiotracers (positron emitters for PET scanning and single photon emitting isotopes for SPECT scanning) are used to label molecules or pharmacologic agents to image molecular interactions of biological processes in vivo. The nanomolar concentration of the radiotracer allows in vivo assessment of biologic processes without interfering with the process itself.
  3. PET and SPECT cameras are able to detect, localize, and quantify regional distribution of radioactivity within the brain. Reconstruction techniques similar to those used in CT (eg, “filtered backprojection” and “iterative reconstruction”) yield 2-dimensional and 3-dimensional images of the brain.
  4. The main advantage of a PET imaging over SPECT imaging is that PET utilizes “coincidence” detection, which is the ability to detect the simultaneous emission of 2 gamma rays that are generated when positrons annihilate when encountering negative electrons. Practically, this results in the greater spatial resolution of PET compared to SPECT. Both techniques can be used to study the same biologic processes.
  5. The biological significance of measured radioactivity on PET and SPECT scans depends on the biological function of the molecule that is attached to the radioisotope. For example, radioligands are available to measure cerebral blood flow, blood brain barrier permeability, neurotransmitter synthesis, enzyme activity, receptor density, glucose metabolic rates, and gene expression, among other physiologic processes. Analysis of regional brain functions can be more specific by performing imaging both before and after specific pharmacological interventions, specialized motor or mental tasks (so-called “activation” studies), or therapeutic interventions, such as stem cell or gene therapy.
  6. PET or SPECT imaging can also be used in the clinical arena for localization of epileptogenic foci, diagnosing dementing disorders, distinguishing between tumor recurrence and radionecrosis, and providing a chronologic record of disease progression (or response to therapy).
  7. Finally, the same SPECT radioligands that are used for cerebral blood flow imaging (such as Tc-99m HMPAO) can also be used for bedside brain death studies using a standard planar gamma camera. Radioligands such as In-111 DTPA can be used intrathecally for radioisotope cisternograms for the clinical evaluation of patients with hydrocephalus or suspected CSF leaks. Radioisotope studies can also be performed to evaluate patency of CSF shunts.

ADVANCED CENTRAL NERVOUS SYSTEM APPLICATIONS OF MRI, CT, AND ANGIOGRAPHY

Magnetic Resonance Imaging

  1. MRA

MR data can be collected using software that images only moving tissues (ie, extravascular stationary soft tissue does not generate any MR signal on these studies while moving blood or CSF will generate signal). Using these techniques, images of blood vessels can be created with a sufficient level of detail that, in many cases, formal endovascular catheter angiography can be avoided (eg, demonstration of significant atherosclerotic disease in the region of the carotid artery bifurcation, surveillance imaging of a known aneurysm, etc). CSF flow evaluation can also be performed (eg, at the craniovertebral junction in Chiari patients).

  1. Diffusion MRI and Perfusion MRI

Using rapidly applied magnetic field gradients, random molecular motion of water molecules can be converted to images and can be presented as “Diffusion” MR images. Rapidly applied gradients can also be used without or with contrast agent administration to assess cerebral blood perfusion.

Diffusion MR imaging is sensitive to early changes of cerebral ischemia, often on the order of minutes. In the setting of acute cerebral ischemia, MR diffusion imaging changes (which usually reflect acute irreversible ischemic change) and MR perfusion imaging changes (which reflect cerebral blood flow) can be performed. The MR perfusion and diffusion images can then be compared. As diffusion images usually represent permanent injury (infarction) and perfusion (blood flow) deficits are potentially reversible, if an MR perfusion deficit is more extensive than an associated MR diffusion deficit, it is possible that an area of perfusion abnormality without diffusion abnormality represents viable brain that is at risk to go on to permanent injury, but is not yet irreversibly injured (the so-called “ischemic penumbra”) and might benefit from aggressive therapy. Alternatively, if a diffusion deficit (representing a region of irreversible brain injury) is similar in extent to a perfusion deficit, then maybe the ischemic changes are permanent and there is no remaining “at risk” penumbra brain, and therefore no aggressive therapy is warranted.

  1. MRS

Using standard MR hardware, MRS can be performed to evaluate brain metabolites. Normal brain tissue includes many metabolites the most important of which are N- acetyl-aspartate [NAA] (which is found in normally functioning neurons and decreases with neuronal injury), Choline [Cho] (a component of cell membranes which increases in any process that increases cellular turnover such as tumors, acute infections, etc) and Creatine [Cr] (a marker of cellular energy). Lactic acid can also be seen in ischemic cells (as a byproduct of anaerobic glycolysis), but is not present in detectable amounts on MRS in normal brain tissue.

Computed Tomography

  1. Biopsy

CT can be used for guidance to percutaneously biopsy many lesions that previously required an open surgical procedure.

  1. Perfusion CT

While rapidly infusing contrast material through a peripheral vein, using a high-speed multidetector CT scanner, data can be collected that reflects how well different portions of the brain are being perfused. In some patients, there are areas of underperfused brain which may manifest clinically with transient symptoms (eg, transient ischemic attacks); these regions might benefit from revascularization before the patient suffers irreversible injury (ie, a stroke). Evaluations similar to diffusion/perfusion brain MRI to assess for penumbra brain (see above) can also be performed.

  1. CT Angiography (CTA)

While rapidly infusing contrast material through a peripheral vein, using a high-speed multidetector CT scanner, data can be collected that can be “reconstructed” with background data removed such that images similar to catheter angiograms can be created without risk of stroke. While intrinsic vascular detail is not as good as endovascular catheter angiography, in many patients, it is adequate to address the relevant clinical issue.

Angiography

  1. Endovascular Coiling of Aneurysms

Small wires can be passed through an endovascular (ie, within a blood vessel) catheter and directly deposited into a saccular aneurysm where they form coils after release from the catheter; an aneurysm can be filled with these wire coils resulting in the complete obliteration of the patency of the aneurysm lumen eliminating the chance that such an aneurysm might rupture in the future and obviating the need for a craniotomy for placement of an aneurysm clip.

  1. Stenting

Instead of open surgery to address a narrowed atherosclerotic or dissected blood vessel, affected blood vessels can be treated via an intravascular catheter which first uses a balloon to dilate the narrowed portion of the blood vessel and then release a mesh stent to maintain the patency of the newly dilated vessel. Stents can also be placed at the base of wide-necked aneurysms allowing for the endovascular coiling of aneurysms that otherwise would not be technically possible.

  1. Treating Intraluminal Thrombus

An acute (within 6-8 hours of onset of symptoms) intraluminal occluding thrombus (clot) or thromboembolus can often be treated with intraarterial catheter placement within the blood clot with off-label (as of 2013) injection of a clot lysing agent (eg, tissue plasminogen activator [t-PA]) to reestablish blood flow to an affected vascular territory before permanent brain damage occurs. Alternatively, some intravascular clots can be captured and removed from the vascular system using specialized catheters [10].

Brinjikji W, Rabinstein AA, Nasr DM, Lanzino G, Kallmes DF, Cloft HJ: Better outcomes with treatment by coiling relative to clippinf of unruptures intracranial aneurysms in the United States. AJNR Am J Neuroradiol 2011;32:1071-1075.

Fleischmann D, Boas FE: Computed tomography—old ideas and new technology. Eur Rad 2011;21:510517.

Hacke W, Kaste M, Bluhmki E, et al: thrombolysis with alteplase 3 to 4.5 hours after acute ischemic stroke. N Eng J Med 2008;359:1317-1329.

Hopyan J, Ciarallo A, Dowlatshahi D, et al: Certainty of stroke diagnosis: incremental benefit with CT perfusion over non-contrast CT and CT angiography. Radiology 2010;255:142-153.

McKinney AM, Palmer CS, Truwit CL, et al: Detection of aneurysms by 64-section multidetector CT angiography in patients acutely suspected of having an intracranial aneurysm and comparison with digital subtraction and 3D rotational angiography. AJNR 2008;29:594-602.

Schellinger PD, Bryan RN, Caplan LR, et al: Evidence-based guideline: The role of diffusion and perfusion MRI for the diagnosis of acute ischemic stroke: Report of the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology. Neurology 2010;75:177-185.

CRANIOCEREBRAL TRAUMA

Hugh J. L. Garton, MD, MHSC Emily Lehmann, MD

OVERVIEW & EPIDEMIOLOGY

Head injury (traumatic brain injury [TBI]) is a leading cause of morbidity and mortality. The incidence of hospitalization for TBI is 75-200/100,000 population. TBI occurs among all ages, peaking in 15- to 24-year-old males. Head injury is very frequent in poly trauma patients managed by trauma surgeons or emergency physicians. Thorough familiarity of the basics of care is therefore highly desirable.

Motor vehicle accidents are the most frequent cause of TBI in the developed world, accounting for 30%-50% of all serious head injuries. Falls and recreational injuries account for about 10%-15% of TBI. Inflicted injury (assault) accounts for about 10%-20% of injuries in adult patients. Age and mechanism or injury are related, as assaults occur mostly to very young infants (child abuse) and to young adults ages 18-24. Falls are the most common source of head injury in patients over 80 years of age. Response to injury also appears to be age dependent. Young people, particularly young men, are more likely to suffer a brain injury, but the chances of dying from that injury are much higher in the elderly. Head injury is the leading cause of death among all patients suffering traumatic injury.

Head injury can be divided on clinical grounds into mild, moderate, and severe forms. About 80% of injuries are mild, including most concussions. Many patients with these injuries do not require hospitalization. Moderate and severe injuries account for about 10% each of the total injury burden and all of these patients are hospitalized. The death rate from head injury is estimated as 20-30/100,000.

Because recovery from moderate and severe injury is often incomplete, survivors of head injury and their care providers often must manage life-long disability.

PATHOPHYSIOLOGY

Classifications/Definitions

Brain injury literature groups injury in several different and unrelated ways which may cause confusion. First, injury can be divided into primary versus secondary types. Primary injury occurs at the time of the brain injury or immediately thereafter. It includes the immediate deformative or concussive forces applied to the brain. Clinically, primary injuries include skull fractures, lacerations of the brain, and hemorrhage in and around the brain (which may take some time to fully accumulate). In contrast to primary brain injury, secondary brain injury denotes the brain’s response to injury, including, for example, loss of regulation of cerebral blood flow, cellular ischemic injury, and brain edema. Most of head injury research involves identification and interruption of these secondary injury pathways. Primary injuries can be induced by either focal or diffuse force application and by either linear or angular changes in momentum. The brain is much more sensitive to a diffuse, angular application of force than a focal, linear one. Experimentally, concussions occur at much lower total acceleration when an angular force is applied compared to a linear one.

Penetrating injury to the brain results in trauma both from the direct disruption of the brain cause by the projectile, as well as compression/decompression injury due to the passage of the bullet or other device. The energy that a projectile imparts to the head is directly proportional to the projectile mass and to the square of the velocity. Velocity therefore is usually the stronger determinant of the extent of projectile injury and the distinction between low- and high-velocity injuries is important.

Secondary brain injury describes the processes that occur in the brain in response to the primary brain injury. These occur from the sub-cellular to the macroscopic level. Calcium-dependent mechanisms, oxidative stresses from free radicals, and apoptotic mechanisms are all likely involved. Tissue ischemia clearly occurs after severe TBI, and in most cases cerebral blood flow (a marker for tissue perfusion) drops considerably in the early phases after severe TBI. The consequence of the microscopic injury cascade at a more macroscopic level is an increase in brain water content: cerebral edema. This occurs both through cytotoxic (cell injury and cell swelling) and vasogenic (incompetent vasculature) mechanisms, although likely the former mechanism is the more important. Brain edema acts as additional mass within the cranial vault that must be accommodated in managing intracranial pressure. The skull forms a rigid, protective covering for the brain. Any increase in the amount of material within the skull, such as a hematoma or edema, must be accommodated within the fixed volume of the skull, and will generally increase the intracranial pressure (ICP) (Figure 36–1). The intracranial compartment is subdivided into compartments by folds in the dura. The tentorium divides the cranial vault into a supra and infratentorial compartment while the falx cerebri divides the supratentorial compartment into right and left halves. This compartmentalization is of benefit after trauma by helping restrict the consequences of injury from impacting the other compartments. However, the compartmentalization and the protections that occur because of it are incomplete and in severe injuries, brain material will herniated out from its compartment of origin often producing specific clinical herniation syndromes.

Clinical Assessment

The basics of management for patients with brain injury can be divided into initial resuscitation, primary neurologic survey, a search for lesions requiring immediate surgical management, and finally, identification and management of cerebral edema and increased ICP. These steps are often applied recursively as the patient condition may change frequently during injury course. The processes required to resuscitate, triage, and manage a concussion are less involved than for a severe brain injury, but the basic steps are similar.

Initial resuscitation for patients suffering brain injury is similar to that for all trauma victims. Management of the airway, breathing, and circulation (the ABCs) is paramount. While patients sustaining isolated mild TBI will usually maintain these functions, patients with multiple injuries, or patients with severe brain injuries often cannot maintain these critical functions without assistance. For example, hypoxia occurs in 30% of patients presenting with severe TBI, who often, because of their injury, cannot protect their airway. In addition to its obvious primary deleterious effects on the brain, hypoxia is also a strong stimulus to increase cerebral blood flow through vasodilatation of the cerebral vasculature. The additional volume of blood in the cerebral vasculature then adds measurably to the ICP after injury. Early endotracheal intubation to secure an airway and provide adequate ventilation is essential. All patients with severe brain injuries and many patients with more serious moderate brain injuries may require intubation, purely on the basis of a reduced ability to protect the airway and maintain ventilation. Intubation of a brain-injured patient should be performed with short acting pharmacologic agents. It is best to avoid the use of long-acting sedating medications and even the reflexive repetitive use of shorter acting agents in the prehospital and trauma bay settings. The loss of the neurological exam that results can impair or delay critical management decisions. Adequate analgesia and sedation are necessary for many trauma patients at intubation and during early resuscitation but the goal should be use the minimum necessary with repeated doses based on reassessment of patient status.

Restoring and maintaining an adequate blood pressure is also of critical importance. This should be accomplished by using intravenous fluids and blood as needed to restore a normal circulating blood volume, and by control of active hemorrhage. The use of pressor agents to support blood pressure may also be appropriate if hypotension persists despite an adequate circulating blood volume. Brain injury, unless it has progressed past a herniation event, is rarely the cause of hypotension, so that a standard search for the source of hemorrhage should accompany resuscitation as for any traumatic injury. Scalp injuries, which often accompany TBI, can be significant source of blood loss especially in children. Database research demonstrates that a single episode of hypotension following a brain injury doubles the risk of dying compared to patients who never have such an episode.

The Glasgow Coma Scale

Once the ABCs have been addressed, the goal of the primary neurologic survey is a rapid, accurate categorization of the severity of the brain injury and a search for clinical evidence of large intracerebral hematomas that could require immediate evacuation. The Glasgow Coma Scale is a three component scoring system that is the main tool of the primary neurological survey (Table 36–1). To apply the scale, a patient is asked to give their name or description of what happened and to follow a simple command. Patients who are unresponsive to verbal requests should receive a brief, firm, noxious stimuli. An effective approach is to apply digital pressure to the trapezius muscle and observe for the response. Additional stimuli may be needed elsewhere to confirm the motor response. A patient can localize when their actions are purposeful and they attend specifically to a noxious stimuli, in an attempt to remove it. A patient is withdrawing when they do not attend to stimuli, but do act to move away from noxious stimuli in a nonstereotypic fashion. Decorticate and decerebrate posturing are stereotypic movements. The possible scores range from 3 to 15. Patients receive the better score when the left and right sides are different. Intubated patients may be given a score of “T” or their verbal score may be estimated from other communicative efforts that they make. The power of this grading scheme is that it is quick and accurate in terms of long term prognosis, assuming the examination is not clouded by sedatives or other confounding factors. Patients who score between 3 and 8 have severe TBI, 9-12 are categorized as moderate and 13-15 as mild closed head injury.

In addition to the determination of the GCS scale, the primary neurological survey includes an assessment of the pupils and gross motor function to assess for the presence of lateralizing signs of a large intracranial hemorrhage. These exams can also alert to the potential presence of a concomitant spinal cord injury. The mass of hematoma, depending on size and location, may produce uncal herniation with unilateral pupillary dilation and contralateral hemiparesis. These findings may indicate that a life threatening intracranial hemorrhage is present, which can be amenable to urgent evacuation. Depending on the situation it may be appropriate to assess other brainstem responses during the primary survey, but more often these are deferred until after an initial head CT scan is obtained. These brainstem reflexes can help to localization injuries (Table 36–3). Two additional brainstem reflexes, the Oculocephalic (doll’s eyes) and Oculovestibular (caloric testing) are not usually tested initial in traumatic settings because of potential exacerbation of cervical spine injuries (doll’s eyes) or basilar skull fractures (caloric testing).

Table 36-3. Common brainstem reflexes evaluated in patients with traumatic brain injury.

Diagnostic Imaging

Once the patient’s airway, breathing, and circulation are safe, the severity of the head injury is determined and the primary neurological survey is completed, all patients with moderate and severe brain injuries should have a CT scan of the head. Rapid use of the head CT is also appropriate in many cases of mild closed head injury, especially when other risk factors, such as the use of anticoagulants, are present. Though the presence of a skull fracture significantly increases the chances of finding a lesion on a head CT, plain skull x-rays are not advised as a substitute to the head CT. MRI may at some point replace CT as the primary imaging tool for TBI, but this has not yet occurred. CT is quicker, safer for management of uncooperative patients, and safer for the environmental issues of loose ferromagnetic materials near, on or in the patient. MRI may be useful in assessing cervical and intracranial vasculature, and it may be helpful in more accurate prognostication of injury outcome.

The head CT is reviewed for the presence of surgically significant intracranial hemorrhage. In addition, there are specific radiographic markers that predict increased ICP (Figure 36–7). These include effacement of the basal and convexity subarachnoid spaces (so-called “cisternal effacement”), mass effect (compression/deformation of adjacent brain structures), and shifting of the brain contents from one side to the other causing “midline shift.” Patients with certain facial, skull base, and cervical fractures are at risk for cervical intracranial vascular injury and dedicated vascular imaging may be indicated. Conventional angiography, CT, and MRI-based angiography are all potential considerations in these situations. Conventional angiography is the gold standard for injury detection and offers the option of endovascular treatment; however, it is also time intensive and may separate the patient from the optimal critical care environment for a protracted period. CT angiography is convenient and rapid, but may miss some injuries. MRI angiography may be more sensitive but has the same downsides as MRI in other trauma settings. No highly regarded predictive tool exists, but risk factors that should prompt consideration of a vascular injury and subsequent vascular imaging include an unexplained neurological deficit, massive facial bleeding, or epistaxis, fracture involving the foraman lacerum of the skull base or foramen transversarium of the cervical vertebrae.

Figure 36–7. A. Normal head CT demonstrating normal basal cisterns (white arrows). B. A head CT demonstrating effacement of basal cisterns (black arrow) and global edema after head trauma.

At the conclusion of the of the head CT scan, the clinical and radiographic information should be synthesized to formulate a plan either to operate to evacuate an intracranial lesion, to evaluate for intracranial hypertension with monitoring technology and clinical examination, or to observe clinically for worsening of neurological status with clinical examinations alone. Radiographic assessments, similar to clinical assessments, are often iterative with repeat imaging used regularly to assess for an alteration in status and optimal management plan.

CLINICAL INJURY PATTERNS/HERNIATION SYNDROMES

There are several important clinical syndromes that herald herniation events within the brain. Each has both an imaging and clinical correlate. Immediate recognition and treatment of these is essential for patient survival. Herniation is the decompensated response to an increasingly large intracranial mass or unchecked cerebral edema. When the ICP or regional compartment pressure reaches a sufficiently high level, the brain tissue is displaced out of the compartment into the adjacent one. If all the intracranial compartments are under equally high pressure, the brain seeks to exit the calvarium through the foramen magnum. Subfalcine herniation occurs when part of the cerebrum is forced from one side to the other under the falx cerebri. This is evident radiographically by the degree of midline shift present on head CT or MRI. Usually the falx itself is shifted to some extent with the brain. In trauma, there is a general correlation between the degree of midline shift, the depression in the level of consciousness, and the severity of the injury. Rarely, the anterior cerebral artery can be pinched by the falx when there is subfalcine herniation. Uncal herniation occurs when the uncus (latin for “hook”) of the temporal lobe is shifted medially by a mass or swelling in the ipsilateral hemisphere. This is most likely to happen with temporal lobe masses because of the close proximity of the mass or swelling to the uncus. Anatomically, the uncus compresses the adjacent subarachnoid space, wherein lies cranial nerve III (oculomotor), resulting in loss of its parasympathetic fiber function to the ipsilateral eye and pupillary dilatation from unopposed sympathetic tone. With more severe or prolonged compression the extraocular muscle function of CN III is lost also, leading to a eye that is deviated interiorly and laterally (unopposed lateral rectus and superior oblique function). If the uncus is forced further medially, it compresses the cerebral peduncle producing contralateral hemiparesis. The clinical syndrome associated with these events therefore is an initial restlessness, then somnolence followed by a dilated ipsilateral pupil, and then contralateral hemiparesis. A somewhat confusing variation is the Kernohan’s notch phenomenon, wherein the entire brainstem is shifted by the uncus and the contralateral peduncle comes into contact with the opposite tentorial edge, leading to weakness that appears clinically on the same side as the pupillary dilation. This is relevant clinically, because if faced with the clinical contradiction of left pupillary dilatation and left hemiparesis, the left pupillary dilation is a stronger predictor of the ipsilateral nature of the cranial problem, than the left hemiparesis which would otherwise predict a right sided problem. Uncal hernation can have the secondary complication of compression of the posterior cerebral artery which runs near CN III. Patients are at risk for a PCA distribution stroke. Tonsillar herniation occurs when the intracranial contents, particular the contents of the posterior fossa, are forced out of the foramen magnum at the base of the skull. The tonsils of the cerebellum are pushed downward and compress the medulla oblongata, leading to respiratory depression and death. Radiographically, this is identified by the loss of CSF spaces around the brainstem and foramen magnum, as brain tissue is pushed into these spaces.

Cushing triad is the constellation of bradycardia, hypertension, and respiratory irregularity that often occupies a herniation event clinically. It is likely due to brainstem compression. The hypertension can be conceptualized as an attempt to protect the brain’s perfusion pressure from the high ICP. Intubation and mechanical ventilation often obscure the respiratory part of the triad, but the other two are observed regularly.

SPECIFIC INJURIES AND SURGICAL MANAGEMENT

Skull fractures are the usually the result of focal application of force to the head. These are usually categorized as open or closed, depressed or nondepressed, and basilar or convexity varieties. Depressed fractures can tear the dura or lacerate the cortex of the brain. Depressed fractures greater than the width of the bone are usually considered for operative repair, particularly if there is an associated laceration of the skin. Skull fractures that are not depressed usually do not usually require repair. Basilar skull fractures (those fractures involving the bones at the base of the brain (parts of the sphenoid, temporal, occipital bones, and clivus) can damage the vasculature and cranial nerves, and can lead to a CSF leak and meningitis. Clinically, one may suspect a basilar skull fracture in the presence of Battles sign, which is retroauricular ecchymosis, or “raccoon eyes,” which is bilateral periorbital ecchymosis. Basilar and nondepressed calvarial vault fractures are usually managed conservatively, although their presence should prompt consideration of vascular and cranial nerve injury. Controversy exists as to the best management of fractures involving the inner table of the frontal sinus, or extending from the skull base into the ethmoid or other skull base sinus. There is the potential for the spread of infection through sinus communication with the epidural space. However, many such fractures heal spontaneously without complications, and management must be individualized.

An Epidural Hematoma (EDH) can occur with a skull fracture that lacerates an artery in the dura (Figure 36–8). The classic example is laceration of the middle meningeal artery by the temporal bone. The bleeding occurs on the outside of the dura and collects in and expands the potential epidural space between bone and dura. Because the source of bleeding is an artery, the hematoma can compress the adjacent brain to the point of serious injury or death. Not all epidural hematomas are caused by an arterial injury in the dura. Bleeding from a skull fracture or dural venous sinus can sometimes collect in the epidural space. These “venous epidural hematomas” are much less likely to produce life-threatening compression of the brain. Distinguishing between the two types is largely a matter of location (temporal vs. nontemporal), size (small vs. large), and rate of change (slow vs. fast). Because the primary injury in an isolated epidural hematoma does not involve the brain, the neurological outcome is excellent if the diagnosis and treatment are prompt. The classic presentation of a temporal epidural hematoma is a patient suffering a blow to the temple, and a brief loss of consciousness. This is followed by a lucid interval during which the patient appears to be neurologically well, as there has been little primary brain injury. During this time, the epidural hematoma either expands slowly enough for the compensatory mechanisms of the brain to maintain consciousness, or the hemorrhage stops temporarily. Subsequently, either due to rebleeding or exhaustion of the compensatory mechanisms, a rapid decline in level of consciousness follows from the mass of hematoma associated with the other elements of uncal herniation, including an ipsilateral CN III Palsy and contralateral hemiparesis. However, this classic sequence is only present in about 25% of patients with EDH. Many patients have no loss of consciousness with injury while conversely, about 20% decline after injury without the lucid interval. Epidural hematomas are usually diagnosed from CT imaging and have a lens-shaped (lentiform) appearance, as the dura is attached firmly to the skull at nearby sutures, tapering each end of the hematoma. CT images are also helpful to judge the impact on the hematoma on the brain, such as the degree of shift of the midline, or compression of adjacent structures. Management of an EDH is determined by lesion size, location, and time from injury to diagnosis. Generally, an EDH more than 1 cm in depth is considered for surgical removal, as are lesions in the temporal fossa where there is less room for further expansion before brain injury occurs. In some cases, an EDH is diagnosed two or more days after the injury that caused it. Since the bulk of hematoma expansion occurs within the first 24-36 hours, it is considered safe to observe patients with EDH of modest size (about 1-cm deep) when the diagnosis is made in a delayed fashion. The hematomas in many of these cases will reabsorb spontaneously. In contrast, a 1-cm-thick temporal epidural hematoma in a patient only 30 minutes out from injury should be strongly considered for emergent surgical removal of the hematoma prior to care of other injuries not immediately threatening to the airway, breathing, or circulation. The operative approach involves positioning the patient to gain access to the site of the hematoma. For a temporal epidural hematoma, a skin incision is made from the root of the zygoma extending into the ipsilateral frontal region in a reverse question mark fashion. As circumstances dictate, a burr hole can be placed in the temporal region, after incising the temporalis muscle before the skin incision is completed to allow blood to escape from the epidural space. However, in most instances, the blood is clotted and rapid completion of craniotomy must follow. The temporalis muscle is elevated and retracted anteriorly. Additional burr holes are placed as needed and a bone flap is elevated. The hematoma is removed. Lacerated dural vessels can be controlled with bipolar cautery or suture depending on size. The dural surface is inspected for an associated subdural hematoma, and intraoperative ultrasound may be useful as needed. The dura is then sutured to the bony margins of the craniotomy to prevent reaccumulation. When a hematoma has collected in the vicinity of a major dural venous sinus, such as the transverse or sigmoid, the sinus itself may be torn and the resulting hemorrhage that is very difficult to control. The outcome for a patient suffering even from a large EDH can be quite good if managed promptly.

Figure 36–8. Acute epidural hematoma after head trauma in a 5-year-old patient.

Acute Subdural Hematomas (ASDH) are the result of trauma to the brain causing rupture of veins over the surface of the brain that transit from the cortical surface to the inner surface of the dura to reach the dural sinuses. When these vessels are injured, the blood collects between the dura matter and the arachnoid membrane, in the potential subdural space. The force required to shear vessels in this fashion can occur with a focal blow to the head, but is likely more common with diffuse, rotational force applications to the brain, as often occur in motor vehicle accidents. Except in the elderly, ASDH generally occurs in conjunction with fairly severe associated brain injury. Focal bruising of the brain is often present beneath the subdural hematoma. Most patients present with a significantly depressed level of consciousness and may have other findings related to the compressive mass of the hematoma. A subdural hematoma has more of a crescent shape to it on imaging studies, because there is no barrier to it spreading over the hemispheric surface of the brain. Associated intracerebral hemorrhage and edema is often evident beneath the ASDH (Figure 36–9).

Figure 36–9. Acute subdural hematoma with an associated intraparenchymal hemorrhage after head trauma in a 25-year-old patient.

Management of the acute subdural hematoma involves emergent operative removal by craniotomy for lesions more than 1-cm thick. A large craniotomy is generally required, and acute brain swelling during the surgical procedure must be expected. Unlike the situation in an EDH, the bleeding source in an ASDH is often difficult to discern. Diffuse hemorrhage arising from under the margins of the bone flap adjacent to dural venous sinuses, is often best managed with gentle packing of hemostatic materials rather than aggressive exploration. Medical management of elevated ICP and the underlying brain injury (see below) is an essential part of the patient’s care. Outcomes are often poor, with a mortality of 50%-90%. Survivors often have significant disabilities. Outcomes can be predicted by the admission Glasgow Coma Scale score and in certain cases medical care is likely to be futile and offered only with reservations.

In the elderly, the atrophy of the brain places the transiting veins under stretch and injury can occur with much less force. Subdural hematomas can present more chronically in such elderly patients, even several months after injury, with a several centimeter thick collection. Such chronic subdural hematomas should be thought of as a different injury from the ASDH (Figure 36–10).

Figure 36–10. Chronic subdural hematoma with acute blood layering in fluid in a 55-year-old patient.

In cases of chronic subdural hematoma, drainage of the hematoma either through a burr hole in the skull or via a craniotomy is indicated. Outcomes are much better than for ASDH, although recurrence and reoperation are common.

Coagulapathy, whether from anticoagulant medications, or as a complication of severe trauma, presents a special risk to TBI patients who are prone to sudden deterioration from hematoma expansion. Immediate, aggressive correction of coagulapathy is indicated for such patients.

Intraparenchymal contusions are common after trauma. These are hemorrhage mixed with brain and usually occur either at the site of a direct blow to the head, or at a point opposite the point of impact. This later phenomena is called a contra-coup injury and is the result of the pressure wave of force moving through the brain and impacting and rebounding from the skull opposite the impact site. Contusions often occur in anterior temporal lobes after frontal impact as the temporal lobes impact into the sphenoid wing anteriorly. The base of the frontal and temporal lobes is also a common site for contusions caused by the brain moving roughly over the irregular bony surfaces of the frontal and middle fossas. Posteriorly, this does not occur as the brain is moving over the smoother tentorium. The clinical presentation of such contusions is specific to their location in the brain, but they can also produce more widespread symptoms through their mass effect as discussed below in secondary brain injury. Intraparenchymal contusions should be distinguished from smaller hemorrhages associated with diffuse axonal injury (DIA) discussed below. There is certainly overlap in terms of size and location, but the typical intraparenchymal hemorrhages are more than 0.5 cm in diameter, and relate to either bony anatomy or force vector as described above. Management for these lesions is generally as conservative as the situation will allow because the hematoma is often intimately mixed with brain, some of which may still be functional. Lesions more than 25 cc are often considered for resection, but the relative eloquence of brain and the response to attempted medical management are often important considerations.

Penetrating brain injury usually produces a focal injury pattern specific to the site of penetration. High-velocity injuries such as those from a gunshot produce, in addition to the focal injury, a large area of cavitation, and hemorrhagic injury from the blast effect. Gunshot wounds that pass through the ventricular system, as a marker for the “middle” of the brain, are most often fatal. Management of less-severe penetrating injury revolves around managing the potential infectious complications of injury and repairing the breach to the skull.

DIFFUSE BRAIN INJURY DIAGNOSES NOT TYPICALLY REQUIRING IMMEDIATE SURGICAL MANAGEMENT

Concussion is an immediate and transient loss of consciousness or normal mentation after head trauma, often associated with a period of amnesia. It is a very common injury. Sport injuries are a common source of concussions and young patients appear to be at highest risk. Concussion is most easily induced by sudden rotation of the head. The presumption is that the cerebral cortex is rotating around the more fixed midbrain and dienchephalon, producing disruption of input and outflow from the reticular activating system. Presentation after concussion can be surprisingly varied, with some patients displaying no loss of consciousness, but rather confusion and amnesia. The severity of concussion appears to be proportional to the duration of amnesia, particularly the anterograde (since injury) amnesia. Relying on consensus/expert opinion, several different criteria for diagnosis and grading systems for severity have been proposed. These systems use the duration of the confusion or amnesia and the presence or absence of a loss of consciousness to form a three-tiered scale of severity and are primarily designed to assist with return-to-play decisions for athletes. To the surgeon, the more typical concern is to decide whether a patient presenting with a concussion needs a CT scan of the head. Validated clinical decision rules stress that all patients with a Glasgow Coma Score of less than 15, those that are vomiting and those that are older than 60-65 years are at high enough risk to warrant a CT scan. Other factors warranting a CT for presenting concussed patients are severe headache, intoxication, persistent anterograde amnesia, a seizure with the injury, evidence of trauma to bone or soft tissue above the clavicle, and severe mechanism of injury such as autopedestrian or ejection injury.

Management for isolated concussions presenting for emergency room evaluation usually includes a period of observation for at least 2 hours, assuming the individual is neurologically normal. Dependent on any other injuries, the patient may be discharged to a responsible adult with written instructions to return for specific symptoms as listed for obtaining a head CT. Patient with abnormal CT scans are usually admitted for care. Long-term outcome is generally good with the majority of patients experiencing no long-term sequellae. However, up to 25% of patients report an increased incidence of headaches and memory difficulties even several months later. Postconcussive syndrome is a more severe version of this phenomenon that includes these and other symptoms such as depression, anxiety, emotional liability, insomnia, and fatigue. Treatment is largely one of reassurance and management of individual symptoms. Because athletes are often concussed, questions arise around when it is reasonable to return to play.

Traumatic Subarachnoid Hemorrhage is a relatively common finding in patients suffering severe TBI. Following injury, a relatively small amount of blood may collect in the subarachnoid space, often over the convexities of the brain, but also in the basal cisternal arachnoid spaces. Trauma is the most common cause of subarachnoid hemorrhage, not rupture of an aneurysm, so the report of subarachnoid hemorrhage after trauma should not automatically prompt a search for a ruptured aneurysm. A history of a neurological collapse before the trauma or more dense blood collecting in the basal cisterns or around the larger intracranial vasculature should prompt more concern for a cerebral aneurysm.

Diffuse Axonal Injury occurs when axons become sheared off at the boundary between gray and white matter during rapid brain acceleration or deceleration. The substance of the cerebral cortex is organized in to a series of alternating layers of gray and white matter (gray cortical mantel, subcortical white matter, deep gray matter nuclei of the basal ganglia, and white matter of the internal capsule). These layers have different tissue densities and when subjected to force in trauma, behave differently. The border between these two tissues is often a site of injury as these two layers accelerate or decelerate at rates according to their tissue properties. DIA is a common finding in severe injury, occurring in up to 50% of pts. Clinically, DIA can occur in varying degrees of severity. In minimal cases, a prolonged mildly concussive state of confusion and memory loss might occur. In more severe cases, the presentation is a depressed level of consciousness. Focal findings can occur if there are DIA hemorrhages in specific locations such as the internal capsule or brainstem. The appearance on head CT is one of multiple small (< 1 cm) hemorrhages, scattered throughout the brain at the junction of gray and white matter. Grading schemes that relate the severity of the CT findings to the eventual neurological outcome exist. Effacement of the basal cisterns and midline shift are examples of radiographic prognosticators of poor outcome after DIA. Management for patients suffering from DIA is primarily medical and focuses on prevention and management of secondary brain injury discussed below.

MEDICAL MANAGEMENT AFTER TBI

The medical management is largely directed at detecting and preventing secondary brain injury from seizures, systemic phenomena such as hypotension and hypoxia and intracranial hypertension.

Seizure Management

The incidence of seizures following TBI is estimated at between 5% and 15%. The majority of these events occur within the first 7 days after trauma. A seizure has the possibility of increasing brain demand for oxygen and nutrients, at a point when the injury may limit the ability of the brain to respond in this fashion. Prophylactic anticonvulsants, usually phenytoin (dilantin), or more recently levetiracetam (Keppra), are reasonably used to prevent seizures after a moderate and severe TBI, when there is evidence of nontrivial brain injury on CT scans. However, experimental evidence only supports their use for the first 7 days after a TBI, and there does not appear to be benefit to continuing the prophylaxis beyond 7 days. Patients experiencing seizures outside of the immediate point of impact are candidates for therapeutic anticonvulsant use extending beyond the 7 day time frame, usually for several months or more, depending on whether the seizures recur.

Homeostatsis

Just as in the initial resuscitation, ongoing care of acute brain injury requires careful protection against hypoxia, hypotension, and hyperthermia. Each of these has the potential to increase ICP or further deprive the brain of adequate glucose or oxygen, or both.

Physiology of Increased Intracranial Pressure

Other than its impact on the neurological examination, secondary brain injury manifests itself clinically as cerebral edema and an increase in ICP. The Monroe–Kelly Doctrine states that given a fixed intracranial volume, consisting of brain, cerebrospinal fluid, and arterial and venous blood, any additional material must be accommodated by a decline in the amounts of the others, or a rise in pressure due to the increase in total intracranial material (Figure 36–1). When a mass is introduced into the cranial vault, CSF and later venous blood are displaced to make room for the mass while initially allowing relatively normal ICPs. However, these mechanisms are overcome and eventually the rise in pressure with increasing volume becomes exponential. The compensatory phase is very important clinically, as a patient may harbor a clinically important lesion while showing only modest symptoms of increased pressure in the brain. This model works well in trauma but is overly simplistic and does not explain clinical phenomena that occur over a longer time frame, such as might occur with chronic hydrocephalus or a brain tumor. In such cases, the compressibility of the brain, usually described as its compliance (change in volume for a given change in pressure) matters a great deal and brain pressures may be normal even in the face of a mass that, if it occurred acutely, would overwhelm the compensatory mechanisms mentioned above. Despite these limitations, the compartment model is very useful for managing cerebral edema and increased ICP. Left unchecked, these consequences of secondary injury lead to increased cerebral edema, increased ICP, compression and finally injury of adjacent brain and its vasculature, producing additional brain ischemia and tissue injury, cyclically generating more brain edema and further increased ICP. ICP is both a surrogate measure of the presence of cerebral edema and also a primary actor in ongoing secondary brain injury. Increased ICP, like hypoxia and hypotension, is a strong predictor of mortality and morbidity after head injury.

The concept of cerebral perfusion pressure (CPP) highlights the ability of elevations in ICP to reduce tissue perfusion:

CPP = MAP − ICP,

where MAP is mean arterial pressure.

In adult patients, CPP of less than 70 mm HG is associated with worsened outcome. The equation also illustrates the importance of maintaining an adequate blood pressure in TBI management. Hypotension is among the strongest predictors of poor neurological outcome after TBI.

Detection of cerebral edema and intracranial hypertension is based on suspicion, radiographic features, and direct measurement of ICP. In severe head injury, as measured by the GCS score, the incidence of increased ICP is 50%-60%. Radiographic findings of concern include the presence of mass effect, midline shift, the loss of the evident pattern of sulci and gyri from displacement of CSF, the loss of distinction of the gray/white matter junction, and the effacement of the basal CSF cisterns. Increased ICP can occur in the absence of any of these findings, so in severely head injured patients, direct measurement of the ICP can still be indicated without concerning radiographic features. At present, it is unproven whether treatment directed by ICP is preferable to one based on CT and clinical examination. Nevertheless current guidelines from the Brain Trauma Foundation recommend ICP monitoring for:

  1. All patients with GCS 3-8 and abnormal head CT;
  2. GCS 3-8 with normal head CT but with hypotension or age more than 40;
  3. Patients in whom the neurological exam cannot be assessed because of sedation or need for general anesthesia, if the suspicion of increased ICP is high.

ICP is measured in trauma by either introduction of a transducer into the brain parenchyma or placement of a catheter into the ventricles of the brain to measure the pressure of the cerebrospinal fluid in a minor surgical procedure. Complication rates of intracranial hemorrhage (2% vs. < 1%) and infection (10% vs. 2%) are higher for catheter-based measurement systems over parenchymal transducers, but only a catheter based system can drain CSF, a distinct treatment advantage. Parenchymal monitors are subject to drift in accuracy and may be inaccurate by 3-4 mm HG after 4-5 days. Catheter-based systems may occlude or be unreliable in the face of very compressed ventricles. In trauma, ICP is not measured by lumbar puncture as there is a risk of brain herniation from higher cranial versus lumbar pressures. Normal ICPs range between about 5 and 15 mm Hg at rest and vary with position and activity. Treatment thresholds after brain injury are typically more than 20 mm HG in adult patients, while children and infants likely require treatment at a lower ICP, although the precise numbers are not yet established. Measurement of the ICP allows calculation of the CPP. Managing brain injury patients based on CPP rather than ICP holds interest as a sufficiently high blood pressure should be able to perfuse the brain despite the ICP. However, clinical research suggests that while CPPs less than 70 mm HG in adult patients are associated with poor outcome, artificially raising blood pressure to supraphysiologic levels in an effort to overcome an increase in ICP, worsens, rather than improves outcomes. CPP measurements currently focus care on avoiding relative hypotension during management.

Treatment options for intracranial hypertension and cerebral edema are most easily understood with reference to the four compartment model of the brain noted above (brain, venous, arterial blood, and CSF). When patients present with elevated ICP, treatments available manipulate the volume of one of these compartments. At present, the selection and style of intracranial hypertension management is as much art as science. Few well-designed studies exist to directly compare different management strategies. Each treatment has risks associated with its use. General strategies include employment of less-risky strategies first and escalation as needed, and to apply methods directed at each of the compartments before applying multiple methods to the same compartment, although many therapies work on multiple parts of the model. The following section covers the treatment options in general but not rigid order of preference. A therapy from the end of the list would very rarely be used before one at the beginning, but that adjacent items in the list might be interchanged.

  1. Improve Venous Drainage

If blood is restricted from exiting the central nervous system, ICP increases as the venous compartment size is larger. In trauma patients, tight fitting cervical collars, a supine body position, and fighting against the ventilator all increase venous pressures. Loosening collars, elevating the head of the bed 30 degrees and minimizing ventilator pressures all improve venous drainage and lower ICPs. These measures have very little risk and can be employed widely.

  1. CSF Drainage

Reducing the size of the CSF space can make more room for brain edema and lower ICP. This is done by draining CSF from the same ventricular catheter used to measure ICP. The risks of this therapy are infection and hemorrhage, and a capable surgeon must be available to place the catheter.

Sedation/Paralysis

Besides caring for the patient’s comfort after an injury, sedation and/or chemical paralysis are important in reducing excessive brain metabolism that accompanies agitation from brain injury. This metabolism can be directly toxic in an injured brain, and obligates increased arterial and venous blood volume to supply the tissue with nutrients. In addition, sedatives/paralytic agents reduce fighting against the ventilator, reducing venous congestion. However, these agents reduce the ability to follow the neurological exam and have the side effect of hypotension when given in excess. Typically, agents such as morphine and ativan are used for analgesia and sedation, while muscular paralytics such as vecuronium are used to facilitate ventilation, but there is little evidence supporting the use of specific regimens with the exception of propofol, which should not be used in children. In general, shorter acting agents are preferable to long acting ones because of the desire to periodically examine the patient without their influence.

Osmotic Agents and Diuretics

In theory, excessive brain water can be removed directly in some cases by establishing a favorable osmotic gradient for diffusion of fluid back into the blood stream. Mannitol, a sugar, and concentrated saline solutions (3% NaCl), do not cross the blood brain barrier and therefore provide such a gradient. Their effect appears to require an intact blood brain barrier. Experimentally, they draw water from less injured areas of the brain, rather than from the more injured areas. This reduces the overall brain volume, and can decrease the ICP, in the four-compartment model. In addition, experimental evidence suggests that a significant part of the ICP reducing effect of osmotic agents comes from reducing the viscosity of blood by altering RBC morphology. This allows delivery of more blood through smaller channels, permitting safe reduction in vessel caliber and, therefore, total intracranial blood volume. The adverse effects of these therapies include dehydration (and hypotension) in the case of mannitol, and nephrotoxicity from increased osmlolarity for all agents. In addition, some of the osmolar particles do make it across the blood brain barrier, and can pull water back into the brain if the therapy is withdrawn too quickly. Dosing regimens for these agents are variable but mannitol is usually used at 0.5-1 gm/kg per dose as often as every 2-3 hours. Dose of 1 gm/kg are given for impending herniation. NaCl 3% may be dosed either continuously at 1-3 cc/kg/h or as bolus doses of similar amounts. Serum sodium and osmolarity measurements should be assessed frequently. For mannitol, a serum osmolarity more than 320 mOsm/L appears to threaten toxicity and limit efficacy, while for hypertonic saline, serum osmolarities of 360 mOsm/dL or more have been reported without renal injury. In this area, the medical literature is limited.

Hyperventilation

The autoregulatory mechanisms of the brain rely in part on CO2 (or perhaps pH) concentrations in the blood. A high metabolic rate leads to increased CO2 production and acidosis. The natural response to this is vasodilatation, to remove waste products and increase the supply of metabolites. Clinically, artificially increasing the respiratory rate (and dropping the CO2) significantly decreases the intracranial blood volume by vasoconstriction. The downside risk is that if done to excess, the vasoconstriction produces ischemia. While hyperventilation was once widely practiced, it is now reserved for situations in which the brain injury has produced an excessive, as opposed to reduced degree of blood flow. This is relatively rare. However, it is import to manage ventilator settings to avoiding elevated CO2 levels (and reduced O2 levels). This avoids unnecessary vasodilation and the increased ICP that results. Arterial CO2 levels of 35 mm HG and O2 levels of 100 mm HG are the common therapy targets.

Barbiturates

High doses of barbiturates reduce cerebral metabolism and experimentally protect against brain injury from the regional ischemia common in secondary brain injury (“the induced coma”). This reduced demand for metabolites decreases the blood flow requirements for adequate cell nutrition, and thus can decrease ICP. However, no clinical experiment has clearly shown that barbiturates improve outcome. The risk is that these agents can profoundly lower blood pressure. They must be used very carefully if at all.

Hypothemia

In experimental settings, hypothermia appears to slow the destructive secondary injury pathways at a cellular level. This reduces the edema that comes from cell death. However, randomized trials in adult head injury have not shown benefit while pediatric trials are ongoing. In the published trial protocols, cooling is typically begun early in the hospital course, if not immediately, with target temperatures of 32-33°C maintained for several days after the initial injury. Complications associated with the therapy in the adult clinical trials have been an increase in infection rate, and serious electrolyte disturbances, particularly hyperkalemia.

Surgical Decompression

Surgical management of cerebral edema involves enlarging the space available for swelling by removing a portion of the calvarium. The surgical technique is to remove a large portion of the skull over the more effected side (hemicraniectomy) or removal of large portions of both frontal bones (bifrontal craniectomy). The dura is generally opened and patched with either allograft or native periosteum. Clinical studies suggest that surgical decompression is effective in lowering ICP, but data regarding neurological outcomes are varied.

OUTCOMES AFTER TRAUMATIC BRAIN INJURY

While many patients with mild traumatic brain injury return to their preinjury level of function, a significant number develop chronic symptoms of fatigue, memory impairment, headaches, and difficulty with concentration. In one study of prospectively followed injury victims, Thornhill, Teasdale, and colleagues reported that over 50% had some identifiable disability one year after injury, including both physical and mental impairments. These were severe enough to impact activities of daily living in one-third to one-fourth of patients. Predictors for poor outcome included age less than 40, and preinjury disability. The medical literature is mixed with regard to the longer term outcome. In a separate study by the Whitnall and coauthors, the overall rates of disability were similar at 1 year and 5 years from injury, but about 25% of patients had exchanged categories from good to disabled and vice-versa. Changes in depression, anxiety, and reported stress appeared to correlate strongly with category change, and only 7% reported the use of rehabilitation services by the 5 year postinjury point. Patients with moderate closed head injuries have more varied outcomes. Most recover to maintain their activities of daily living and even return to work or school. Detailed neurocognitive testing, however, often reveals deficits in executive function and memory. There are more consistent reports of chronic fatigue and headaches. Severe TBI is often a life altering event for both patient and family event. The prognostic implications of injury are often vitally important for family members making decisions about what degree of aggressive care to provide. As an average, perhaps 15%-20% of all patients with severe injury will make a good recovery, while more than 50% will either die or be severely disabled. Sadly, for some patients, the degree of injury and the poor likelihood of meaningful recovery make the provision of aggressive care an exercise in futility. Elderly patients, those more than 80 years old, with severe head injuries have a very poor prognosis. For patients with the worst prognostic features, such as older age, very low GCS or 3-5 without improvement with resuscitation, lack of pupillary response to light and associated chest or abdominal injuries with hypotension, the prognosis for meaningful recovery is very poor. Once the diagnosis and injury severity are confirmed by examination and imaging studies and the condition is unchanged despite resuscitation and withdrawal of all pharmacologic agents likely to be affecting the examination, a gentle but frank discussion of the situation is necessary and appropriate. However, generally, the younger the patient is, and the higher the presenting GCS score, particularly the motor GCS scores, even within the severe injury group, the better the chances of some degree of recovery. The physician would do well to remember that prognostic information represents a probability of an outcome not a certainty of it. While some families may appreciate knowing these details, others will more appreciate whatever kernels of hope can be provided under the circumstances.

Bullock MR, et al: Surgical management of traumatic parenchymal lesions. Neurosurgery 2006;58(3 Suppl):S25-S46; discussion Si-iv.

Guidelines for the management of severe traumatic brain injury. J Neurotrauma 2007;24(Suppl):1.

Thornhill S, Teasdale GM, Murray GD, McEwen J, Roy CW, Penny KI: Disability in young people and adults one year after head injury: prospective cohort study. BMJ 2000;320(7250):1631-1635.

Timofeev I, Czosnyka M, Nortje J, Smielewski P, Kirkpatrick P, Gupta A, Hutchinson P: Effect of decompressive craniectomy on intracranial pressure and cerebrospinal compensation following traumatic brain injury. J Neurosurg 2008;108(1):66-73.

Wakai A, McCabe A, Roberts I, Schierhout G: Mannitol for acute traumatic brain injury. Cochrane Database Syst Rev 2013 Aug 5;8:CD001049.

Whitnall L, McMillan TM, Murray GD, Teasdale GM: Disability in young people and adults after head injury: 5-7 year follow up of a prospective cohort study. J Neurol Neurosurg Psychiatry2006;77(5):640-645.

Wood RL: Long-term outcome of serious traumatic brain injury. Eur J Anaesthesiol 2008;42(Suppl):115-122.

SPINAL CORD INJURY

John Ziewacz, Frank La Marca

General Considerations

Traumatic spinal cord injury (SCI) is devastating. It primarily affects young people, and often results in significant disability or death. The median age at diagnosis of SCI is 37.6 years. The main causes in order of incidence are: motor vehicle collision, fall, violence, and sports injuries. Despite maximal medical and surgical therapy, the prognosis for significant recovery of a complete lesion is poor. Recent research into stem cell technology and other new modalities of therapy have not yet been effective in human clinical trials.

The cost to the healthcare system and to society is significant. A 25 years old with a high cervical cord injury (C1-4) is estimated to incur $741,425 in medical costs in the first year following SCI, and $132,807 for each year survived thereafter. In addition, the loss of wages and productivity for SCI patients averages $57,000 annually. With 12,000-14,000 Americans suffering SCI per year, the social and economic costs are significant.

The demographics of SCI have changed in the last 30 years. The median age of 37.6 years has increased from 28.7 years in the 1970s. This is a largely due to an increase in the incidence of falls causing SCI in patients more than 60 years of age. Though the relative incidence of both sports injuries and violent injuries has declined over the last 30 years, the larger decrease in sports injuries has placed it below violent injuries as a cause of SCI.

Treatment for SCI consists of acute and chronic management. Acutely, the ABCs (airway, breathing, circulation) must be secured, and the spine immobilized to prevent extension of injuries. Compressive lesions must be identified and the need for urgent surgical management determined. Methylprednisolone is currently a treatment option used widely in the initial phases of SCI, though it is associated with side-effects which may sometimes outweigh the potential benefits of its use. Other acute nonsurgical treatment options aimed primarily at minimizing secondary injury are modest induced hypothermia and hyperbaric therapy. The first would be based on the neuroprotective properties of hypothermia in brain injury but has not yet shown to have proven beneficial effects in humans with traumatic SCI. The second has shown to increase speed of neurological recovery but not an overall improvement in final outcome. Chronic management includes physical, and occupational therapy designed to maximize functionality. Specific therapy and improvement depends on the level and completeness of the injury.

Clinical Findings

Clinical findings in SCI depend on the level, mechanism, and severity of injury. Injuries can be classified as either complete or incomplete. A complete SCI refers to the lack of motor or sensory function below the level of the lesion. Incomplete lesions spare some degree of sensory and/or motor function below the level of the lesion. Incomplete lesions often result in recognized SCI syndromes based on the region of the spinal cord affected. The American Spinal Injury Association publishes a scale to further classify the severity of SCI (Table 36–4).

Table 36-4. The ASIA classification of spinal cord injury.1

A = Complete: No sensory or motor function preserved in the lowest sacral segments (S4/5).

B = Sensory incomplete: Sensory but no motor function preserved below the neurologic level including the sacral segments S4/5.

C = Motor incomplete: Motor function is preserved below the neurologic level, and more than half of the key muscles below the neurologic level have a muscle grade less than 3. There must be some sparing of sensory and/or motor function in the segments S4/5.

D = Motor incomplete: Motor function is preserved below the neurologic level, and more than half the key muscles below the neurologic level have a muscle grade greater than or equal to 3. There must be some sparing of sensory and/or motor function in the segments S4/5.

E = Normal: Sensory and motor functions are normal. Patient may have abnormalities on reflex examination.

1Adapted from: American Spinal Injury Association. Standards for neurological classification of spinal cord injury (revised 2000). Chicago: ASIA; 2002.

Initial clinical findings include motor deficit, sensory deficit, and hyporeflexia. Initially, all reflexes below the lesion are lost including the bulbocavernosus, cremasteric, and abdominal cutaneous reflex. Over time, these reflexes may return, and the deep tendon reflexes become hyperreflexive due to the loss of descending tonic inhibition of the reflex arc. Initially, paralysis is flaccid, but eventually upper motor neuron signs develop, and a spastic paralysis results. If the lesion is in the high cervical region (C1-5), respiratory effort may be compromised due to the loss of innervation to the phrenic nerve. Loss of bowel or bladder function often occurs, and loss of rectal tone and sensation, as well as priapism may result. The loss of bladder control manifests as urinary retention, and developing urinary incontinence is typically overflow incontinence. Loss of anal sphincter tone and sensation results in leakage of stool and lack of awareness of bowel movements.

An important early finding that can occur in SCI is “spinal shock.” This refers to a drop in the systolic blood pressure with accompanying bradycardia, often to a level of 80 mm Hg systolic following SCI. This is due to the loss of sympathetic tone to the regions below the lesion and causes venous pooling and decreased venous return to the heart.

Chronic clinical findings in SCI are related to the long-term need for ventilatory support, immobilization, and need for catheterization. Pneumonia, urinary tract infections, and decubitus ulcers are common findings, and are often the cause of death in spinal cord injured patients.

Incomplete SCIs may demonstrate a variable pattern of sensory or motor preservation, though they can often be categorized into recognizable clinical syndromes, depending on the mechanism of injury and the portion of the cord affected.

  1. Central Cord Syndrome

Central cord syndrome (CCS) refers to a pattern of injury that affects the motor strength in the upper extremities more severely than the lower extremities. Sensory function is variable below the level of the lesion, and sphincter control is often affected. This usually occurs in older patients with spinal stenosis following a hyperextension injury. The central cervical cord is a watershed vascular territory that is thought to be disrupted in this syndrome. The spinal cord is somatotopically organized such that cervical fibers are more medial compared to fibers traveling to the lower extremities, resulting in the more severely affected upper extremities.

  1. Anterior Cord Syndrome

Anterior cord syndrome (ACS) results from the compression of the anterior portion of the cord by a herniated disk, bone fragment, or from occlusion of the anterior spinal artery. The corticospinal tracts and spinothalamic tracts are preferentially affected due to their more anterior location. The posterior columns are relatively spared. This results in loss of motor function and loss of pain and temperature sensation below the level of the lesion, with preserved proprioception, vibration, and pressure sensation. It is important to distinguish surgical from nonsurgical (ie, anterior spinal artery occlusion) etiologies in this condition.

  1. Brown–Séquard Syndrome

Brown–Séquard syndrome occurs after spinal cord hemisection. It is usually the result of penetrating trauma occurring in 2%-4% of spinal cord injuries. Motor function and posterior column function (proprioception, vibration sense) is disrupted on the side of the lesion. Pain and temperature sensation is diminished on the contralateral side due to the crossing of the spino-thalamic tract in the spinal cord at or one or two levels above the entrance of the fibers into the cord.

  1. Conus Medullaris Syndrome

Conus medullaris syndrome (CMS) results from injury to the sacral spinal cord. Symptoms include saddle anesthesia, loss of bowel/bladder function, and lower extremity weakness. It includes a combination of both upper and motor neuron signs.

  1. Cauda–Equina Syndrome

Cauda–Equina syndrome refers to compression and dysfunction of the lumbo-sacral nerve roots. It is not a true SCI as it only affects the nerve roots and not the cord itself. The clinical syndrome is similar to CMS with saddle anesthesia, loss of bowel/bladder function, and lower extremity weakness, but findings are all lower motor neuron.

Physical Examination

Initial physical examination in SCI focuses on the ABCs (airway, breathing, circulation). The spine should be immobilized to prevent further injury. Special attention must be focused on the airway in high cervical injuries as patients may require endotracheal intubation given injury to the nervous supply to the diaphragm. Blood pressure must be closely monitored given the possibility of spinal shock. This manifests as a drop in the systolic blood pressure and must be addressed immediately in order to prevent further cord ischemia.

In the awake patient after the ABCs have been attended to, a history focusing on mechanism of injury and detailed neurologic examination is undertaken in order to determine the level and completeness of the injury. Motor strength should be tested in all muscle groups and sensation should be tested with pinprick, and proprioception. Rectal tone and sensation should be tested with digital examination. Reflexes should be examined including the bulbocavernosus, cremasteric, and abdominal cutaneous reflexes. Careful palpation of the spine is important to evaluate for obvious step-offs or tenderness to palpation at all levels. The examination must be carefully documented and a neurologic level and evaluation of completeness of the lesion determined.

In the comatose patient, a complete neurologic examination is often difficult. In this situation, observation of spontaneous movements or movements to painful stimuli is important. Deep tendon reflexes should be examined and palpation of the spine should be undertaken to observe for obvious step-offs. Radiologic imaging is often required to adequately determine a level of injury and its etiology.

Differential Diagnosis

After a complete history and physical examination is performed, with the addition of radiographic imaging, the diagnosis of SCI is usually apparent. Radiographic imaging can help determine the mechanism of the injury, which is usually due to a fracture or subluxation of the bony spinal elements.

Some peripheral nerve lesions may resemble SCI, but these can usually be distinguished after careful examination and knowledge of the anatomy of the spinal cord and the peripheral nervous system. Peripheral injuries are typically unilateral and affect only lower motor neurons. Sometimes malingering and conversion disorders may mimic SCI. Serial examinations and inconsistencies in examination, in the setting of unremarkable imaging usually permits differentiation.

Radiologic Examination

In the asymptomatic patient with no spinal tenderness, no distracting injury (eg, long-bone fracture), and no evidence of disturbed consciousness or intoxication, no radiographic imaging is necessary. Patients with spine tenderness, numbness, tingling, or obvious signs of SCI (eg, weakness, loss of bowel/bladder control) require radiographic imaging. The hallmark of radiographic imaging has been three-view cervical spine x-rays with AP/Lateral films of the thoracic and lumbar spine. Current recommendations of the American Association of Neurological Surgeons/Congress of Neurological Surgeons (AANS/CNS) recommend three-view cervical spine x-rays in conjunction with CT scanning of the cervical spine in patients with suspected SCI. With the advent of CT scanning with detailed coronal and sagittal reconstructions, CT scanning alone has replaced x-ray examination as the initial diagnostic study of choice in many centers (Figure 36–11). This obviates the need for multiple x-rays and diagnostic/treatment delay in the case of inadequate plain films. MRI is listed as an option in the diagnosis of SCI by the AANS/CNS as it can better detect ligamentous/soft tissue injury, though it often “overcalls” injuries that do not cause instability and may lead to prolonged and unnecessary immobilization. MRI is often reserved for patients for whom SCI signs and symptoms are present and no clear evidence is found on x-ray imaging or CT, or a herniated disk or other soft-tissue abnormality is suspected. MRI (particularly T2-weighted sequences) also clearly demonstrates compression of and/or signal change within the spinal cord (Figure 36–12).

Figure 36–11. Sagittal reconstruction of a cervical computed tomography (CT) scan demonstrating a C6-7 traumatic fracture.

Figure 36–12. Sagittal T2-weighted magnetic resonance image of a C6-7 traumatic fracture demonstrating spinal cord compression and signal change within the spinal cord (arrow).

In awake patients, clearance of the cervical spine consists of normal x-rays (or CT w/reconstructions), CT scan, and flexion/extension views or MRI obtained within 48 hours of injury. At this point, cervical immobilization may be discontinued.

Under current recommendations in obtunded patients, cervical spine clearance may be obtained following normal x-rays, CT, and dynamic flexion/extension films performed under fluoroscopic guidance, normal MRI obtained within 48 hours of injury, or at the discretion of the treating physician. However, with the advent of CT with sagittal and coronal reconstruction providing greater sensitivity for the detection of injury, and the propensity of MRI to “overcall” injuries, some centers clear the cervical spine in obtunded patient with normal x-rays and CT scans. Clearance of the cervical spine in this population is still cause for debate and recommendations are in flux. Further studies will elucidate the necessary and sufficient studies to definitively clear the cervical spine in this population.

In the patient with SCI the combination of x-rays with CT scanning has high sensitivity and identifies the vast majority of lesions causing SCI.

Treatment

Initial treatment of SCI consists of securing the ABCs and immobilizing the spinal column. In the case of high cervical spine injury, the need for endotracheal intubation must be identified and if it is not immediately necessary serial arterial blood gas assessments should be monitored to evaluate for hypocapnia and progressive ventilatory failure. Spinal shock and the resultant decrease in blood pressure must be treated aggressively if it occurs. Volume expansion should be initiated promptly and decreased systolic blood pressure refractory to volume expansion should be treated with pressor therapy. The choice of pressors has not been conclusively defined, but typically a beta-agonist is followed by an alpha-agonist given the possibility of bradycardia in spinal shock.

Patients should be placed in a hard cervical collar and cervical immobilization should be ensured until clearance of the cervical spine, or definitive treatment has occurred. Patients should be placed on a board for transfers and log-rolled for movement until the thoracic and lumbar spine is cleared.

Other initial management considerations include placement of an arterial line to monitor blood pressure on a constant basis, and placement of a Foley catheter to decompress the bladder.

Methylprednisolone has been used in the acute phase of SCI based on studies that demonstrated motor improvement in patient groups that received methylprednisolone in the early period after SCI. However, due to the lack of demonstrated clinical significance of any improvement, and studies demonstrating side-effects of high-dose methylprednisolone, it is offered as an option by the AANS/CNS current guidelines with the knowledge that “evidence suggesting harmful side-effects is more consistent than any suggestion of clinical benefit.”

Following initial stabilization and imaging studies, the need for surgical intervention is assessed. Surgery has two main goals: decompression and stabilization. Surgery is employed on an emergent basis for incomplete lesions in the hopes of preserving or improving neurologic function, and on a nonemergent basis for complete lesions, as there is no demonstrated improvement in neurologic function for emergent surgery for complete lesions. Goals of surgery in this setting are to prevent cranial extension of injury, and to prevent progressive deformity. Choice of surgical approach for SCI is not standardized and is dependent on the location of the pathology. Current stabilization procedures typically involve instrumented fusion techniques, and may be approached via an anterior, posterior, or combined approach.

Cervical traction may be employed either alone, or as an adjunct to surgical therapy to attempt realignment of the spinal column. This is accomplished by fixing a halo ring, or specialized devices (ie, Gardner–Wells tongs) to the head, connecting this to a rope and pulley system, and adding weight to adjust the spine in the desired vector.

Chronic treatment of SCI focuses on rehabilitation and adaptation to permanent injury. Rehabilitation can often result in improved neurologic function in incomplete lesions and can help those with complete injuries become as functional as possible. Patients may require ventilatory support, tracheostomy, intermittent catheterization, frequent turning (to prevent decubitus ulcers), and functional accommodations such as wheelchairs and other devices aimed at improving functionality. Attention to long-term care issues can prolong the life and productivity of SCI patients.

Prognosis/Outcome

Despite exciting research into novel treatments, SCI remains a devastating injury. Death in the acute trauma setting from SCI is 20%. Complete lesions that remain so at 72 hours are unlikely to improve beyond one level above the lesion in the long term. Patients with quadriplegia who have initial ventilator dependency have 5-year survival rates of approximately 33%. Incomplete lesions have a more favorable outcome. Among recognized SCI syndromes, CCS and BSS have the most favorable outcomes, with up to 90% of BSS and CCS patients being able to ambulate independently at 1 year. ACS patients have a worse prognosis, with 10%-20% recovering functional motor control. Causes of death in long-term SCI patients are usually due to cardiac, respiratory, or infectious causes—often related to the sequelae of SCI.

Current multidisciplinary approaches to SCI, including emergency department, medical, surgical, and rehabilitation staff provide the best therapy for patients with SCI. Despite this, SCI remains a devastating injury with high rates of mortality and permanent disability. Novel research and innovations will hopefully provide better outcomes for SCI in the future.

American Spinal Injury Association: Standards for neurological classification of spinal cord injury (revised 2000). Chicago: ASIA; 2002.

Kwon BK, Mann C, Sohn HM, et al: Hypothermia for spinal cord injury. Spine J 2008 Nov-Dec;8(6):859-874.

Ho CH, Wuermser LA, Priebe MM, Chiodo AE, Scelza WM, Kirshblum SC: Spinal cord injury medicine. 1. Epidemiology and classification. Arch Phys Med Rehabil 2007;88(Suppl 1):S49-S54.

McKinley W, Santos K, Meade M, Brooke K: Incidence and outcomes of spinal cord injury clinical syndromes. J Spinal Cord Med 2007;30:215-224.

Priebe MM, Chiodo AE, Scelza WM, Kirshblum SC, Wuermser LA, Ho CH: Spinal cord injury medicine 6. Economic and societal issues in spinal cord injury. Arch Phys Med Rehabil 2007;88(Suppl 1):S84-S88.

PERIPHERAL NERVE LESIONS

Cheerag Upadhyaya, MD Linda Yang, MD John McGillicuddy, MD

General Considerations

Familiarity with the pertinent aspects of peripheral nerve anatomy and physiology combined with focused history and physical examination aids in the management of peripheral nerve lesions. The history and physical examination may then be complimented by electrodiagnostic and radiographic studies.

Microscopic Anatomy

Peripheral nerves are composed of varying combinations of sensory and motor axons. An axon is a long projection from a nerve cell body that is bounded by a cell membrane as well as a basement membrane. Some axons are surrounded by sheaths of myelin, a fatty substance secreted by Schwann cells. Myelin insulates the axon, thereby increasing the velocity of neurotransmission. The axon is, in turn, surrounded by a layer of connective tissue called endoneurium. Axons travel together in bundles called fascicles, each of which is covered by another layer of connective tissue called perineurium. Fascicles are grouped together to form a peripheral nerve, which is surrounded by a final layer of connective tissue called epineurium.

Gross Anatomy

A peripheral nerve is comprised of fibers from more than one spinal nerve root and each spinal nerve root contributes fibers to more than one peripheral nerve. Spinal nerve lesions manifest as radiculopathies with blurred sensory disturbances, while peripheral nerve lesions demonstrate sharply demarcated sensory disturbances. Spinal nerve lesions results in mild to moderate weakness in muscles supplied by one spinal nerve, but by more than one peripheral nerve. Peripheral nerve lesions manifest more severe muscle atrophy and weakness in muscles supplied solely by the peripheral nerve.

ACUTE PERIPHERAL NERVE INJURIES

General Considerations

Common etiologies of acute peripheral nerve injury include penetrating trauma, blunt trauma, traction, fractured bones, or compression from hematomas. Minor peripheral nerve injuries arise from blunt trauma that temporarily compresses or stretches a nerve, but leaves its axons intact (neurapraxia). In such cases, axonal transport may be temporarily impaired, but Wallerian degeneration, or the death of axons distal to the point of injury, does not occur. These injuries generally recover spontaneously over the course of days to weeks. Slightly more severe injuries may interrupt axons and their myelin sheaths while leaving endoneurium intact (axonotmesis). In these cases, Wallerian degeneration inevitably follows. Axonal regeneration may occur spontaneously, however, guided to areas of previous innervation by intact endoneurial tubes. With this type of injury, there is a good prognosis for spontaneous functional recovery, with axonal regeneration occurring at a rate of about 1 mm/d or 1 inch/mo.

Peripheral nerves may be severed cleanly (neurotmesis), as in the case of iatrogenic scalpel injuries during surgery. If the divided ends of the nerve remain in proximity, regeneration can occur via axonal sprouting from the proximal stump. These axonal sprouts may bridge the gap to the distal stump, propagating through preserved endoneurial tubes at a rate of 1 mm/d. Severe crush injuries may create internal damage to a peripheral nerve without completely transecting it. Such injuries disrupt axons and their endoneurium and disturb the organization of fascicles within the nerve. In these cases, fibrous scar tissue may form within the macerated nerve which can block the regeneration of axonal sprouts. A tangle of axonal sprouts contained in fibrous scar tissue is called a neuroma. Neuroma formation acts as a barrier to spontaneous peripheral nerve regeneration.

Clinical Findings

A careful clinical history and a meticulous neurological examination are paramount in determining which peripheral nerves have been injured and the type of injury present. The type of trauma will generally suggest whether or not the nerve is in continuity. A penetrating injury with a sharp object, such as a knife, suggests a clean transection that is amenable to immediate surgical repair. Nonpenetrating trauma or a stretch injury is more suggestive of nerve continuity.

Determine the timing of motor and sensory deficits may also aid in the assessment of the nerve injury. For example, in the setting of a penetrating sharp injury, an immediate deficit at the time of the injury would suggest direct involvement of the peripheral. However, a delayed deficit would suggest an enlarging adjacent lesion such as a hematoma or pseudoaneurysm.

The physical examination includes inspection, observation, evaluation of the relevant vasculature, range of motion assessment, and neurological examination. A laceration, fracture, or bruising/abrasions may suggest the site of an underlying nerve injury. A Horner’s sign (ptosis, meiosis, anhydrosis) is suggestive of a proximal T1/lower trunk brachial plexus lesion. An impaired range of motion can make assessment of strength difficult to evaluate accurately. An elevated hemidiaphragm is suggestive of a phrenic nerve injury.

The sensory and motor findings associated with acute peripheral nerve injury vary widely, depending on which particular nerve is injured. Pain may also be a symptom, but it usually develops in a delayed fashion. Pain can occur as a result of neuroma formation, where it is often associated with a tender lump in the area of injury. Neurogenic pain may also develop because of a disturbance in the processing of pain signals. This type of pain, when associated with autonomic hyperfunction, is referred to as complex region pain syndrome (formally known as causalgia or reflex sympathetic dystrophy); it is notoriously difficult to treat. When neurogenic pain is associated with nerve root avulsion it is known as deafferentation pain. Deafferentation pain often responds well to surgical intervention via dorsal root entry zone ablation.

In the diagnosis of acute peripheral nerve injury, EMG and nerve conduction studies are generally not useful until at least three weeks after injury. Nevertheless, it is important to obtain baseline electrodiagnostic studies as they are important for monitoring recovery. In the case of brachial plexus injury, it is useful to obtain an MRI scan or CT myelogram to look for pseudomeningocoeles in the vicinity of the nerve roots, which would indicate nerve root avulsion.

Certain peripheral nerve injuries are associated with traumatic fractures of specific bones. For example, the radial nerve is particularly vulnerable to injury from fractures of the humerus. Traumatic injuries of the radial nerve classically occur with fractures of the shaft of the humerus, at the level of the spiral groove. Such injuries result in weakness of wrist extension, finger extension, and thumb extension, as well as numbness over the radial aspect of the dorsal surface of the hand. In this type of injury, elbow extension is not affected, since muscular branches to the triceps are given off proximal to the spiral groove.

Trauma to the brachial plexus can cause a wide array of neurological signs and symptoms. Clinical manifestations are determined by the location of the lesion within the brachial plexus as well as the severity of the injury. Erb-Duchenne palsy is a well-described condition involving injury primarily to the upper trunk of the brachial plexus (derived from C5 and C6 nerve roots). It typically results from a stretch injury such as traction on the arm at the time of birth, or a fall that forcefully separates the head from the shoulder. The resulting deficits to the deltoid, biceps, rhomboids, brachioradialis, supraspinatus, and infraspinatus, leave the arm hanging to the side, internally rotated and extended at the elbow. This posture is often called the “waiter’s tip position.”

Differential Diagnosis

In acute trauma, when unilateral limb findings are present, it is important to differentiate acute radiculopathy from peripheral nerve injury. A thorough neurological exam is critical. Several general principles should be considered. Radiculopathy is often accompanied by neck or back pain, which tends to radiate down an arm or a leg. Also, the sensory findings of radiculopathy tend to be blurred, reflecting the overlapping nature of dermatomes, while sensory findings in peripheral nerve injuries are sharply demarcated. Weakness from radiculopathy occurs in muscles innervated by one spinal nerve, but by more than one peripheral nerve. Thus, it is often only partial weakness, since nearly all muscles are innervated by more than one spinal nerve.

One crucial task in diagnosing acute peripheral nerve trauma is to rule out ongoing neural compression. Acute trauma to a peripheral nerve usually results in maximal deficits at the time of injury. A peripheral nerve deficit that progresses should raise a red flag and initiate further workup. Immediate surgical exploration should be considered in order to address compressive lesions such as expanding hematomas or growing traumatic pseudoaneurysms. Sources of ongoing neurologic compression should be removed as soon as possible.

Treatment and Prognosis

Sharp nerve transections with clean ends (knife wounds, for example) should be repaired within three days. The repair should be performed in an end-to-end fashion, with no tension across the repair site. Transections from penetrating trauma that do not have clean edges or where significant tissue loss is present, should be repaired in a delayed fashion. During exploration of the wound, transected nerve stumps should be identified and tagged. The tagged nerves ends should be attached to fascia to reduce the possibility of retraction. After three weeks, the lesion should be reexplored and the injured nerve repaired. At that time, areas of axonal damage and neuroma formation are easier to visualize. Better visualization of damaged axons reduces the possibility of subsequent neuroma formation at the site of repair.

In the case of nonpenetrating injury, where stretch or temporary compression is the likely etiology of the problem, recovery often occurs spontaneously, without the need for surgery. In these cases, nonoperative management with serial neurological examinations, including electrophysiologic testing, should be the initial treatment modality. If, after three months, there is no sign of clinical recovery, the injured nerve should be explored. Intraoperative electrophysiologic nerve mapping should be performed to determine if there is conduction across the site of injury. If there is no conduction of evoked potentials, the neuroma should be resected. The stubs should be trimmed and brought together primarily if it can be done without creating tension. If primary anastamoses of the two nerve ends would result in tension across the repair, then a nerve graft should be employed (usually the sural nerve). If intraoperative stimulation reveals conduction across an area of injury, than the nerve should be left intact and allowed to regenerate on its own. A well-known mnemonic for remembering the appropriate timing of operative repair for traumatic nerve injuries is “the rule of threes”: three days for a sharp transection, three weeks for a open ragged transection, and three months for a closed stretch injury.

Peripheral nerve repair is generally performed under the microscope using 8-0 or 9-0 suture for coaptation. Many surgeons now use tissue glue, rather than suture, to coapt the nerve endings. In the case of nerve root avulsions, which cannot be repaired directly, nerve-transfer procedures may be employed, such as coapting a fascicle from an intact ulnar nerve to a nonfunctioning musculocutaneous nerve in order to restore elbow flexion.

Prognosis for recovery depends on the type of injury as well as the treatment. Axonal regeneration occurs at a rate of one inch per month, proximally to distally. Thus, clinical recovery may proceed slowly. Maximal recovery occurs over the course of approximately one to two years. Rehabilitation and physical therapy is important for avoiding the development of muscle contractures that may limit mobility when nerve function has returned. Tendon transfers may be of assistance if neural function does not completely recover.

PERIPHERAL ENTRAPMENT NEUROPATHIES

General Considerations

Peripheral nerves are subjected to chronic mechanical forces, such as compression, stretching, and friction. These forces, when applied over time, may lead to peripheral nerve entrapment syndromes, such as carpal tunnel syndrome, or ulnar nerve entrapment. Both static and dynamic factors may contribute to chronic peripheral nerve injury. Static factors include musculotendinous anomalies or inflexibleanatomic tunnels that compress peripheral nerves. Dynamic factors include mobile joints, muscular contraction, or nerve mobility that leads to stretching of a peripheral nerve or increased friction along its course during movement. Peripheral entrapment neuropathies occur more frequently in upper rather than lower limbs, probably because of the greater mobility of the arms.

Clinical Findings

Entrapment neuropathies are characterized by weakened muscles as well as sensory disturbances in the distribution of a single peripheral nerve. In general, any given individual muscle is supplied by one peripheral nerve. Thus, entrapment of a peripheral nerve can lead to severe motor findings in a muscle it innervates. Muscle atrophy and fasciculations are not uncommon in peripheral nerve entrapment. Marked atrophy on clinical examination should raise the surgeon’s suspicion that a peripheral nerve lesion is present. Sensory complaints associated with entrapment neuropathy generally include parasthesias, rather than pain, in the distribution of the involved nerve. Percussion over the nerve may result in an electric sensation radiating along the nerve and its territory. In the clinical setting, this is known as Tinel’s sign. Electrodiagnostic testing is also useful in the diagnosis of peripheral nerve entrapment. The finding of a nerve conduction delay at the site of compression on nerve conduction studies is common to all compression neuropathies.

The most common peripheral nerve entrapment syndrome is median nerve entrapment at the wrist. This condition generally arises from compression of the median nerve by the transverse carpal ligament, and is therefore referred to as carpal tunnel syndrome. Carpal tunnel syndrome occurs with higher frequency in patients with conditions leading to connective tissue thickening: rheumatoid arthritis, acromegaly, hypothyroidism, pregnancy, and may be related to repetitive hand or wrist movements. Common presenting symptoms include dysesthetic pain in the hands that is worse at night, often awaking the patient from sleep. This occurs because many people sleep with flexed wrists, a position which exacerbates median nerve compression at the wrist. The pain sometimes radiates upward, into the forearm. Patients also complain of numbness on the palmer side of the hand as well as the first three to three and a half digits, including the thumb. When the ring finger is involved, the sensory disturbance “splits” the finger, involving the radial side of the digit only. The intrinsic hand muscles innervated by the median nerve in the hand are sometimes referred to as the “LOAF” muscles, stemming from a commonly used mnemonic device: Lumbricals (first and second only), Opponens pollicus, Abductor pollicis brevis, and Flexor pollicis brevis. Patients with carpal tunnel syndrome may complain of decreased grip strength, or difficulty grasping small objects. Atrophy of the abductor pollicus brevis, at the lateral base of the thumb may be present. There is often a positive Tinel’s sign at the wrist. Phalen’s test is a clinical maneuver that is sometimes used in the diagnosis of carpal tunnel syndrome. The patient’s wrists are held in forced flexion for at least thirty seconds. The test is considered positive if this maneuver produces symptoms of median neuropathy in the hand.

Ulnar nerve entrapment is the second most common peripheral nerve entrapment syndrome, behind carpal tunnel syndrome. The most common site of ulnar nerve entrapment is at the elbow. Patients typically present with upper extremity pain that localizes to the medial aspect of the elbow. Paresthesias and numbness of the small finger and the ulnar half of the ring finger are also common. A Tinel’s sign is often present at the medial aspect of the elbow: tapping the patient over this area sends electric sensations into the fourth and fifth digits. The ulnar nerve innervates most of the intrinsic hand muscles, including the adductor pollicis, the first dorsal interosseous, and the hypothenar muscles. Patients complain of hand weakness, leading to reduced grip strength and pinch strength. They complain of dropping things, or of trouble opening jars. Atrophy of hand intrinsic muscles may be marked, particularly the first dorsal interosseous and the muscles of the hypothenar eminence. When nerve dysfunction is severe, the hand may take on a “claw hand” appearance. The patient may also exhibit a Froment sign: when asked to hold a piece of paper between the thumb and index finger, the distal interphalangeal joint will flex because the patient fires the flexor pollicis longus muscle, innervated by the median nerve, to compensate for lack of abductor pollicis function.

Differential Diagnosis

The differential diagnosis for peripheral nerve dysfunction includes neuropathies of infectious origin (both bacterial and viral), hereditary conditions (such as Charcot–Marie–Tooth disease), neuropathy associated with nutritional deficiency (such as vitamin B12 deficiency), metabolic or endocrinologic conditions (such as diabetes), inflammatory or immune-mediated conditions (such as polyarteritis nodosa), and toxic conditions (such as lead poisoning.) When more than one peripheral nerve is involved, the surgeon should be wary of the diagnosis of generalized neuropathy, which is typically symmetrical and bilateral. Decreased amplitude on electrodiagnostic studies (which suggests axonal loss) is characteristic of neuropathy of hereditary or metabolic origin. Peripheral nerve entrapment, on the other hand, causes damage to the myelin surrounding an axon, thereby slowing conduction velocity but not affecting amplitude.

Chronic cervical or lumbar radiculopathy must also be ruled out. Several features that separate radiculopathy from peripheral nerve dysfunction were outlined in the section on acute peripheral nerve injuries. It is important to recognize that sensory changes that “split” the ring finger suggest ulnar nerve dysfunction, rather than C8 radiculopathy.

Treatment and Prognosis

Surgical management of peripheral nerve entrapment generally involves decompressing the involved nerve. In the case of carpal tunnel syndrome, the transverse carpal ligament is divided, thus relieving compression on the median nerve as it passes from the wrist into the hand. In the case of ulnar neuropathy, simply freeing the nerve from surrounding scar tissue or hypertrophied connective tissue (external neurolysis) is usually sufficient.

Surgical management of peripheral nerve entrapment should be strongly considered when patients present with significant motor weakness or muscle atrophy. Surgical intervention should also be considered when patients fail to improve with medical treatments. Nonoperative management strategies include avoidance of repetitive activities that precipitate symptoms, or the use of immobilization braces that hold joints in positions that avoid compression.

Carpal tunnel release results in excellent relief of symptoms in 80% of patients and partial relief in another 10%. Similar results are observed for surgical management of ulnar neuropathy.

PERIPHERAL NERVE TUMORS

General Considerations

Peripheral nerve tumors can be divided into non-neoplastic masses, benign masses, and malignant masses. The non-neoplastic masses include traumatic neuromas, morton neuromas, and nerve sheath ganglion cysts. Benign masses include neurofibromas, schwannomas (also know as neurilemmoma), perineuriomas, lipofibromatous hamartoma (also know as neural fibrolipoma), nerve sheath myxoma, and finally granular cell tumors. Malignant masses include the malignant peripheral nerve sheath tumor (MPNST).

Neurofibromas can be subdivided into solitary, diffuse, and plexiform varieties. The solitary neurofibroma is the most common benign peripheral nerve tumor. Axons are incorporated within the neurofibroma along with Schwann cells, collagen matrix, perineurial cells, and fibroblasts. Because the axons are intermixed within the tumor, the tumor cannot be excised without resecting a portion of the involved nerve. Diffuse neurofibromas and plexiform neurofibromas are less common than the solitary variety. The diffuse neurofibroma typically involves the skin and subcutaneous tissues. Plexiform neurofibromas are large, irregular expansions of nerves ranging from small cutaneous nerves to large trunks. Neurofibromas are commonly found in patients with neurofibromatosis Type 1 (NF1/von Recklinghausen’s disease).

Schwannomas are slowly growing lesions made up of Schwann cells in a collagen matrix. Differentiating schwannomas from neurofibromas is that the nerve fascicles run alongside the tumor, rather than through the tumor as in neurofibromas. Schwannomas are the second most common peripheral nerve tumor. Generally, only mild neurologic deficits occur, and operative resection can be considered in the setting of pain, paresthesias, or weakness in the distribution of the nerve.

MPNST is the most common malignant peripheral nerve tumor. Nearly two-thirds arise from neurofibromas in the setting of NF1. The remainder arise either de novo or in patients with history of prior external beam radiation.

Clinical Findings

The symptoms of peripheral nerve tumors are generally related to the nerve involved and include weakness, numbness, paresthesias, and pain. MRI with gadolinium contrast can be useful in imaging the tumors.

Differential Diagnosis

The differential diagnosis of peripheral nerve tumors include entrapment neuropathies, non-neoplastic masses, radiculopathies, and neuropathies associated with infection, malnutrition, metabolic, endocrinologic, inflammatory, immune-mediated, and toxic conditions. Generally, a symptomatic peripheral nerve tumor will be palpable. Further, the peripheral nerve tumor can occur anywhere along the course of the nerve, whereas entrapment neuropathies typically occur in defined locations and the various other neuropathies listed are often diffuse processes. Finally, a radiculopathy will be symptomatic in the distribution of the nerve root, likely involving several peripheral nerves.

Treatment & Prognosis

Schwannomas can generally be resected without any neurologic deficit since the nerve fascicles run alongside the tumor. The fascicles can generally be dissected free. Operative resection is curative. Neurofibromas generally cannot be resected without neurologic deficit since the axons run within the substance of the tumor. Indications for surgical resection of solitary neurofibromas are large tumor mass, accelerated growth, neurologic deficit, and pain. Plexiform neurofibromas (excluding superficial ones) can rarely be totally resected and should generally be followed.

Generally, the prognosis for benign peripheral nerve sheath tumors is very good with improvement in pain, weakness, and paresthesias noted after resection of solitary lesions. MPNST require aggressive surgical treatment. The expected 5 year survival is 15%-20% in patients with NF-1 and 56% in patients with de novo disease.

O’Brien M on behalf of the Guarantors of Brain: Aids to the Examination of the Peripheral Nervous System. 5th Ed. Edinburgh: Saunders Elsevier; 2010.

Song JW, Waljee JF, Burns PB, et al: An outcome study for ulnar neuropathy at the elbow: a multicenter study by the surgery for ulnar nerve (SUN) study group. Neurosurgery 2013;72(6):971-981.

BRAIN TUMORS

Daniel Orringer, MD Shawn Hervey-Jumper, MD

Clinical Presentation

Whenever possible, the evaluation of a brain tumor patient begins with a detailed history focusing on the most common symptoms observed in patients bearing intracranial mass lesions. Headaches are the most common symptom in brain tumor patients, occurring in at least 50% of patients at some point. Classically, brain tumor patients present with headaches that are worse upon waking in the morning and may often be severe enough to wake a patient from sleep. This type of headache is thought to occur as a result of temporary increases in intracranial pressure caused by physiologic elevations in PCO2common during sleep. Normally, transient increases in PCO2 that occur during sleep do not result in headache. However, in the brain tumor patient, during sleep, the combination of elevated intracranial pressure due to mass effect from the presence of the tumor and cerebral vasodilation due to increased PCO2, causes an additive increases in intracranial pressure, ultimately resulting in headache. Headaches related to elevated intracranial pressure can also occur throughout the day as a result of maneuvers that raise intracranial pressure, such as straining, coughing, or bending over.

More commonly, headaches seen in brain tumor patients are not related to elevations in intracranial pressure. Headaches typically caused by brain tumors occur as the neoplastic process involves intracranial structures containing pain fibers. Unlike the brain parenchyma, the dura is richly innervated with pain fibers and is likely the most common source of headaches in brain tumor patients. Dural irritation is also involved in the pathogenesis of other types of headaches. This pathophysiologic overlap may explain the clinical observation that brain tumor-associated headaches may lack any distinguishing characteristics. Brain tumor patients often describe their headaches as deep and aching but these features are highly variable. These may have been previously misdiagnosed as sinus, tension, or migraine headaches. Headaches in brain tumor patients may also be due to visual difficulties in the setting of direct or indirect (through elevated intracranial pressure) tumor effects on the optic pathways and oculomotor nerves.

A number of attributes of the headaches associated with brain tumors can provide useful diagnostic information. First, brain tumors in patients who present with headaches are more likely to be found in noneloquent or functionally silent areas of the central nervous system. Second, headaches occur more frequently in patients with rapidly growing brain tumors. Rapidly growing brain tumors commonly cause severe headaches from meningeal irritation, hemorrhage within the tumor, and/or obstructive hydrocephalus. Obstructive hydrocephalus is a neurosurgical emergency that must be considered in any brain tumor patient presenting with the acute onset of severe headache. In addition, the location of headaches in brain tumor patients commonly provides localizing information. Brain tumors are usually located ipsilateral to the most severe pain.

Seizures are another common presenting finding in patients with brain tumors. Interestingly, seizures are more common with low-grade gliomas than high-grade gliomas. The incidence of seizures in patients with low-grade glioma is estimated as high as 85%. Seizures are the initial presenting symptom in 9% of patients with metastatic brain tumors and 18% of patients with high-grade glioma. Moreover, 25%-50% of all patients with brain tumors experience seizures at some point in their disease course. While seizures associated with brain tumors can be disabling, they can also lead to early diagnosis, and, therefore early treatment.

The nature of seizure activity may hold diagnostic significance. Subcortical and cortical tumors are more likely to cause seizures than those of deeper structures. Focal seizures with primarily motor phenomena, such as tonic-clonic seizures, often occur as a result of involvement of tumor with the primary motor cortex within the frontal lobe. Temporal tumors resulting in temporal lobe seizures typically have variable manifestations that may make localization difficult. Focal seizures due to parietal lesions may present with language disturbance, somatosensory abnormalities, or vestibular symptoms. Focal seizures caused by brain tumors may be secondarily generalized and may ultimately affect multiple cortical regions, causing variable symptomatology. Status epilepticus also can occur as a presentation of brain tumors.

Syncope, another common presentation of brain tumor, must be distinguished from seizure. There are multiple pathophysiologic mechanisms underlying syncope in brain tumor patients. In patients with tumor burden resulting in chronically elevated intracranial pressure and decreased brain compliance, a sudden additional rise in intracranial pressure may compromise cerebral blood flow, resulting in syncope. Transient increases in intracranial pressure that may result from sneezing, coughing, vomiting, or straining are tolerated in patients with normal physiology but may cause syncope in brain tumor patients. Syncope caused by transient increases in intracranial pressure may represent impending herniation and requires urgent neurosurgical attention.

In addition, there are a number of symptoms that commonly present in association with headache, seizure, or syncope. Nausea and vomiting are present at the initial encounter with at least 40% of patients with brain tumors. Nausea and vomiting may be caused by elevations in intracranial pressure and/or direct tumor involvement of the area postrema in the dorsal surface of the fourth ventricle.

Cognitive decline is common especially in the elderly patient and is often misdiagnosed as Alzheimer’s disease. Cognitive decline may easily be confused with depression and is thought to result from generalized fatigue, loss of appetite, and interest in everyday activities. Frontal tumors frontal masses are commonly associated with cognitive decline. Frontal masses, especially those affecting both frontal lobes may also result in apraxia and urinary retention.

A key component of the interview of the brain tumor patient the past medical and family history. The incidence of central nervous system metastases is increasing due to improved survival in patients with the most common types of solid organ cancers. A family history of brain tumors may suggest a familial cancer syndrome. Among the most common familial syndromes predisposing to brain tumor occurrence include von Hippel–Lindau syndrome, tuberous sclerosis, neurofibromatosis 1 and 2, Turcot syndrome (familial adenomatous polyposis), and Lynch syndrome (hereditary nonpolyposis colorectal cancer).

Physical Findings

Physical findings are highly variable depending on tumor location and extent of disease. Nonspecific physical findings can occur with elevated intracranial pressure including papilledema (edema of the head of the optic nerve associated with engorgement of retinal veins) and oculomotor palsy (due to uncal herniation). However, the most helpful physical findings are those that assist in localizing the lesion (Table 36–5). Focal neurologic signs such as muscle weakness are common and when caused by peritumoral edema may be rapidly reversible with the administration of steroids.

Table 36-5. Brain tumor-localizing signs and symptoms.

Aphasia suggests involvement with cortical language centers located in the dominant frontal or parietal lobe. Aphasic patients may be misdiagnosed with dementia or psychiatric disorders. The diagnosis of brain tumor should be considered in, patients without psychiatric history who develop a psychiatric disorder.

Imaging

Radiographic imaging is performed to confirm the clinical diagnosis of brain tumor. Imaging provides information regarding localization, tumor type and the effect of a lesion on surrounding structures. Due to its wide availability, speed, and affordability, noncontrast CT is commonly the initial screening test for patients with brain tumors. CT is also the test of choice for evaluating the extent of tumor invasion into adjacent bony structures. CT angiography can be helpful in evaluating blood supply to tumors or in evaluating the relationship of blood vessels to the tumor.

Whenever possible MRI of the brain with and without gadolinium-based contrast is performed in the brain tumor patient. Traditional morphologic MRI is performed to assess tumor location, size, cellularity, associated cystic components, associated edema or hemorrhage, necrosis, margins, and invasion into surrounding structures, vascularity, and enhancement. Morphologic data can be used to estimate the WHO grade and suggest the tissue diagnosis of a lesion. However, the gold standard for brain tumor diagnosis remains tissue histology. High-quality contrast MRI images are vital for defining the relationship of tumor tissue to eloquent cortical areas and, consequently an operative plan. In addition, MRI images can be reconstructed to create three-dimensional models that can be used during surgery.

Metabolic MRI or magnetic resonance spectroscopy (MRS) can be used to supplement information obtained from traditional morphologic MRI. MRS is used to compare the small molecule content of tumor tissue and normal surrounding brain tissue. MRS improves the accuracy of brain tumor diagnosis by differentiating brain tumors from lesions that appear similar on routine MRI such as abscesses. In addition MRS detects subtle changes in small molecule content that correlate with tumor grade. In treated brain tumors, MRS enables differentiation between radiation necrosis and residual tumor.

A number of alternative magnetic resonance techniques have clinical application in the imaging of brain tumors. These can enable clinicians to make more accurate preoperative diagnoses and provide information about interaction of tumor tissue with adjacent functional cortical structures. Diffusion MRI characterizes brain tumors based on measurement of molecular mobility and is useful in differentiating tumors from similar-appearing lesions, estimating cellularity, and measuring response to treatment. Perfusion MRI is useful for evaluating tumor angiogenesis, endothelial permeability, and response to treatment. Functional MRI maps functional cortical areas and can be used to create an operative corridor or plan for resection that minimizes risk to eloquent surrounding structures. Similarly diffusion tensor imaging defines the integrity of white matter tracts surrounding a tumor and is commonly used in the planning of both surgical and radiation therapy.

Traditional catheter-based cerebral angiography has both historical and contemporary significance in brain tumor imaging. Cerebral angiography was once used to infer tumor location and morphology by measuring displacement of blood vessels by a tumor. Currently, angiography is used in the context of highly vascular lesions, including some meningiomas and hemangiomas, for preoperative embolization. Embolization of vascular tumors diminishes operative risk and difficulty.

Tumor Types

  1. Gliomas

Fifty percent of newly diagnosed brain tumors are primary tumors of glial origin (astrocytes and oligodendrocytes). Glial tumors are stratified in a scale of increasing aggressiveness. Grades one and two are classified as low-grade gliomas whereas, grades three and four are high grade. High-grade gliomas are faster growing and consequently carry a worse prognosis than low-grade gliomas.

Low-Grade Gliomas: Astrocytes, Oligodendrogliomas, and Mixed Gliomas—Approximately 26% of newly diagnosed glial tumors are astrocytomas and 2% are oligodendrogliomas. Between 1500 and 1800 new low-grade gliomas are diagnosed in the United States each year. WHO grade 1 gliomas are reserved for pilocytic tumors. Pilocytic astrocytomas represent 5.2% of all primary intracranial tumors in adults and 20% of all brain tumors in children younger than 15 years. WHO grade 2 lesions are diagnosed based on their infiltration and tendency to progress to higher grade lesions over time. The most common subtypes of low-grade gliomas include juvenile astrocytomas, diffuse astrocytomas, oligodendrogliomas (representing 6.5%), and mixed gliomas.

The etiology of low-grade gliomas is unknown. Genetic studies suggest that mutation or deletion of the tumor suppressor gene TP53 plays a role in the tumorigenesis of low-grade gliomas. Known oncogenic signaling pathways have consequences on tumor metabolism promoting a cellular switch to aerobic glycolysis. Isocitrate dehydrogenase 1 and 2 (IDH 1 and 2) catalyze the decarboxyation of isocitrate into alpha ketoglutarate. IDH1 and IDH2 mutations are found in 40% glioma (70% of low grade, 50% of grade III, and 5%-10% of primary glioblastoma). The impact of these mutations on low-grade diffuse gliomas remains unclear; however, they offer a robust and independent survival benefit in all tumor grades.

Low-grade astrocytomas occur with peak incidence in the young adult population (most commonly in 20-40s). They originate in white matter regions within the CNS, grow slowly and distort surrounding brain structures. Histologically, there is a modest increase in cellularity, disruption of the normal orderly pattern of glial cells, and elongated nuclei. There is no endothelial proliferation or tissue necrosis. Three histologic subtypes of low-grade astrocytomas include fibrillary, gemistocytic, and protoplasmic.

Oligodendrogliomas occur predominantly within the gray matter of the cerebral hemispheres, are well-circumscribed, calcified, and have a slight predominance for the frontal lobes. Like astrocytomas, they occur predominantly in younger patients with most frequent diagnosis in the third decade of life. Histologically, oligodendrogliomas are characterized by uniform cell density and round nuclei withperinuclear halos appearing as a classic “fried egg” appearance. Oligodendrogliomas rarely show a mutation in TP53. In 1994 a codeletion in the long arm of chromosome 1p36 and the short arm of chromosome 19q13 was shown to predict chemosensitivity and better prognosis.

Radiographically, low-grade gliomas are iso- or hypodense to brain on CT scan and do not enhance with contrast. Calcification is common in oligodendrogliomas. On MRI, low-grade glioma are iso- to hypointense on T1-weighted imaging (T1WI) and typically hyperintense on T2-weighted imaging (T2WI) and are not contrast-enhancing.

High-Grade Gliomas—The term malignant glioma includes anaplastic astrocytoma (AA), glioblastoma multiforme (GBM), gliosarcoma, and malignant oligodendroglioma (Figure 36–13). There is wide difference in the prognosis, aggressiveness, and response to therapy between the different tumors in this group.

Figure 36–13. Magnetic resonance imaging of a deep right temporal mass proven to be a high-grade glioma. (A) FLAIR imaging demonstrating a mass with significant surrounding white matter edema. (B) T1-weighted contrast enchanced image demonstrating smaller enhancing portion of the mass.

Malignant astrocytoma, the most common type of adult brain tumor, makes up 15% of all intracranial tumors and 50%-60% of primary brain tumors. While relatively rare, malignant astrocytoma is the fourth most common cause of cancer-related deaths. The incidence of AAs and GBM increases with increasing age. There is little difference in incidence from nation to nation, however, in the United States these tumors are less common among Africans and African Americans.

The majority of malignant gliomas occur sporadically. However, patients with the autosomal recessively inherited, Turcot syndrome have a high rate of malignant glioma (usually medulloblastomas and astrocytomas) in combination with familial adenomatous polyposis. Similarly, patients with tuberous sclerosis and neurofibromatosis type 1 and 2 often develop brain tumors including gliomas.

The most significantly mutated genes in glioblastoma include TP53 (seen in 42% of patients), PTEN (seen in 33%), neurofibromatosis 1 (NF1) (21%), EGFR (18%), PIK3R1 (10%), and PIK3CA (7%). Mutations of TP53 have been identified in the autosomally dominantly inherited Li–Fraumeni syndrome, which results in malignant gliomas in addition to tumors involving breast, blood, bone, and the adrenal cortex. It has recently been suggested that a subpopulation of cells with stem-like properties (cancer stem cells) are present in glioblastoma offering resistance to chemotherapy and radiation.

A hallmark of malignant gliomas is the propensity to invade and migrate along white matter tracts. Invasion increases with increased grade and growth factors such as epidermal growth factor increase this invasion. Autopsy studies show that malignant glioma cells spread through the CSF and extend into and beyond areas on MRI with T2 signal change. Histologically, grade 3 gliomas show mitotic activity and nuclear atypia but no necrosis, while grade 4 tumors have nuclear atypia, mitoses, endothelial proliferation, and areas of necrosis. The radiologic hallmarks of GBM are ring enhancement and areas of central necrosis detected on CT and MRI.

Gangliogliomas—Gangliogliomas are rare tumors found most commonly in patients between the ages of 15 and 20 with a history of seizures. They represent 1% of CNS neoplasms in adults, 7.6% in children. They can be found in any region of the central nervous system, but seem to occur predominantly in the temporal lobes. Gangliogliomas are distinguished histologically from pure gliomas by their mixture of neuronal and glial elements. Calcification is common. Macroscopically, they may appear solid or cystic. They are generally well-circumscribed, cystic tumors that may display a mural nodule projecting into the cyst cavity. Imaging characteristics vary in their enhancement and signal qualities on MRI. Lesions can exhibit cystic or solid components or a combination of both. Tumor calcification is a common imaging feature. Gangliogliomas are typically benign WHO I or II tumors with indolent behavior offering a 93%-98% 5-year survival. Five percent of gangliogliomas, however, are aggressive anaplastic or malignant gangliogliomas WHO grade III-IV based on the presence of increased cellularity, microvascular proliferation, and areas of necrosis.

Brainstem Gliomas—Brainstem gliomas represent 10%-20% of all CNS tumors in children. Brainstem gliomas are a heterogeneous group with diverse clinical presentations, prognoses, and patterns of growth. They are described as focal, diffuse, cervicomedullary, and dorsally exophytic. Focal tumors are less than 2 cm in size with a well-circumscribed appearance on MRI and no surrounding edema. They are most prevalent in the midbrain and medulla, but can occur at any level in the brainstem. These children typically present with focal cranial nerve deficits and contralateral hemiparesis. Diffuse tumors (diffuse intrinsic pontine gliomas) account for the majority of brainstem gliomas (80%) and commonly arise in the pons. These patients typically present with bilateral cranial nerve deficits, ataxia, and long tract signs. Cervicomedullary brainstem gliomas take their origin from the upper cervical cord and extend rostrally into the cervicomedullary junction and often present with lower cranial nerve deficits and long tract signs. Dorsal exophytic tumors account for 20% of brainstem gliomas and arise from the floor of the fourth ventricle. They are typically sharply delineated from surrounding structures. These patients present with cranial nerve deficits, elevated intracranial pressure, and failure to thrive.

MRI scan has allowed for the recognition of the four classes of brainstem glioma. Even though MRI contrast signal poorly correlates with histologic grade, MRI provides adequate anatomic visualization. Focal tumors are classically well circumscribed and small without infiltration or a significant amount of surrounding edema. Dorsal exophytic tumors arise in the floor of the fourth ventricle and are typically hyopintense on T1-weighted imaging, hyperintense on T2-weighted imaging, and homogenously enhance with gadalinium contrast. Diffuse brainstem tumors are hypointense on T1-weighted images and hyperintense on T2 sequences. Because MRI characteristics for diffuse tumors are highly specific an accurate diagnosis can be made in the majority of cases. Recent literature has shown that the mortality associated with biopsy of brainstem tumors may have been modestly exaggerated. Biopsy is therefore considered in cases of abnormal clinical presentation or imaging.

  1. Primitive Neuroectodermal Tumors

Primitive neuroectodermal tumors (PNET) are thought to originate in cells from primitive neural crest. PNET includes medulloblastomas, pinealoblastomas, ependymoblastomas, esthesioneuroblastomas, and neuroblastomas. PNET are more common in children than adults. Medulloblastomas are PNETs within the posterior fossa and account for 20% of childhood brain tumors and 1% of all adult tumors. Medulloblastomas are the most common primary central nervous system tumor in children younger than 18 years old.

Several syndromes result in an increased incidence of medulloblastomas, including tuberous sclerosis, neurofibromatosis, Gorlin syndrome, and Turcot syndrome. Loss of portions of chromosome 17 either through deletions or unbalanced translocation is associated with over 50% of medulloblastomas. Over the past decade, transcriptional profiling of medulloblastomas revealed the existence of 4 distinct subgroups; WNT, SHH, Group 3, and Group 4. WNT medulloblastomas have a classic histology, WNT gene expression signature, and the best prognosis with more than 95% in 5 years. These patients tend to be the least common (10% of cases) and are rarely metastatic. SHH-driven medulloblastomas exhibit a desmoplastic histology (although occasionally large cell or anaplastic are seen). Patients SHH tumors represent an intermediate prognosis with survival ranging from 60% to 80%. Group 3 medulloblastoma have the worst prognosis with 50% metastatic at the time of diagnosis. These tumors exhibit aberrant MYC expression with focal high-level amplifications. Group 4 medulloblastomas account for 40% of cases with an intermediate prognosis similar to the SHH subgroup. Group 4 tumors are driven by the oncogenes MYCN and CDK6 (cyclin-dependent kinase 6). Unlike other subgroups group 4 medulloblastoma predominately affect men and metastases are seen in 30% of cases. These subgroups have reestablished what was previously considered a single tumor entity but are now requiring different therapeutic approaches.

Grossly, medulloblastomas typically occur within the cerebellar vermis and are poorly demarcated, purplish, soft, and friable. Histologically, these tumors are highly cellular, composed of homogenous fields of small, round, blue-cell tumors with hyperchromatic nuclei, minimal cytoplasm, and occasional calcification. Other histologic characteristics of these tumors include varying degrees of neuronal and glial differentiation, Homer Wright rosettes (nuclei surround a clear central area of cell processes indicative of neuroblastic differentiation) are often present and mitotic figures are numerous.

Meduloblastomas are PNET within the posterior fossa. Histologically similar tumors within the pineal gland are pinealoblastomas, and within the supratentorial space are neuroblastomas. Retinoblastomas are histologically similar tumors within the eye, PNET originating from olfactory epithelium are termed esthesioneuroblastomas, and intraventricular PNET are ependymoblastomas.

On CT, medulloblastomas are typically hyperdense, homogenously enhancing, and occasionally cystic. Small areas of calcification can be appreciated on CT. Scattered areas of hemorrhage, necrosis, and calcification can occur. On MRI, medullolastomas are isointense or hyopintense to brain on T1WI, hyperintense to brain on T2WI and intensely contrast enhancing. If medulloblastoma is suspected, MRI of the spine is obtained to rule out metastases. Lumbar puncture should be performed with extreme caution because the majority of children have associated obstructive hydrocephalus.

  1. Pineal Tumors

The pineal gland is bounded ventrally by the quadrigeminal plate and midbrain tectum, dorsally by the splenium of corpus callosum, rostrally by the posterior aspect of the third ventricle, and caudally by the cerebellar vermis. Tumors in this region are usually found incidentally on MRI and are most common in children, making up 3%-8% of pediatric brain tumors. The pineal region has several diverse cell types including glial cells, arachnoid cells, pineal glandular tissue, ependymal lining, sympathetic nerves, germ cells, and remnants of ectoderm. Tumors within this area can therefore be grouped into 4 categories: germ cell tumors, pineal parenchymal cell tumors, glial cell tumors, and other miscellaneous tumors and cysts. In the pediatric population germinomas and astrocytomas are the most common tumor type. Germ cell tumors and pineal cell tumors occur primarily during childhood. In the adult population, pineal tumors are more commonly gliomas and meningiomas.

Germ cell tumors, ependymomas, and pineal cell tumors can metastasize through the CSF causing myelopathic or radiculopathic symptoms. Pineal tumors typically present with symptoms of increased intracranial pressure from obstructive hydrocephalus, direct brainstem and cerebellar compression and endocrine dysfunction. In addition, Parinaud syndrome (upgaze paralysis, convergence-retraction nystagmus, pseudo-Argyll Robertson pupils, eyelid retraction, and conjugate downgaze in the primary position) is associated with pineal tumors.

MRI is the primary diagnostic imaging modality for pineal tumors, but it does not reliably predict tumor histology. In contrast, tumor markers may be useful in the diagnostic process, to determine response to treatment, or as an indicator of early recurrence. Elevation of serum or CSF alpha-fetoprotein (AFP) or human chorionic gonadotropin (HCG) suggests a germ cell tumor. Mildly elevated AFP suggests the presence of a fetal yolk sac tumor. Marked elevation of AFP suggests endodermal sinus tumors while smaller elevations are suggestive of embryonal cell carcinoma or immature teratoma. HCG is often markedly elevated with choriocarcinomas.

  1. Ependymoma

Ependymomas were previously thought to arise from the ependymal cells lining the ventricles and central canal of the spinal cord; however, they have recently been shown that radial glial cells are the cells of origin. They occur in both children and adults and 65% occur within the posterior fossa (most commonly in children). Ependymomas are quite rare representing only 6% of all gliomas in adults. However, they are the third most common brain tumor in children (behind pilocytic astrocytomas and medulloblastomas). Three cases per 100,000 children younger than 15 are diagnosed each year with this tumor type.

The etiology of posterior fossa ependymomas is unknown however significant advancements have been made over the past decade on their biological profile allowing the identification of different molecular subgroups. Group A patients have laterally located tumors and younger patients with more than 50% recurrence rate (independent of extent of surgical resection). Group B patients have a slightly more favorable prognosis and older age at diagnosis. Familial cases have also been identified. As with several other primary CNS tumors, these tumors often have loss of heterozygosity of chromosome 22q, which contains the neurofibromatosis 2 (NF2) gene. Patients with neurofibromatosis have an increased incidence of gliomas including ependymomas.

A histologic grading system that correlates with tumor aggressiveness is used to classify ependymomas. Histologically, ependymomas are often characterized by, epithelium-like cells in a rosette pattern, formed by a ring of polygonal cells surrounding a central cavity. Tumors may also exhibit perivascular pseudorosettes, intranuclear inclusions, calcifications, and papillary clusters.

The imaging characteristics of ependymomas are variable, but they are typically isodense to cerebral cortex on noncontrast head CT. Calcifications and cystic components within the tumor are frequent. On MRI the solid portion is typically isointense to gray matter on T1WI and isointense to hyperintense on T2WI.

  1. Cerebral Lymphoma

Cerebral lymphoma involving the brain, spinal cord, or ocular structures can occur either primarily or as a metastasis. The source of premalignant lymphocytes is controversial because the CNS lacks lymphoid tissue. CNS lymphoma occurs most commonly in severely immunocompromised patients where it is commonly associated with Epstein–Barr virus infections. CNS lymphoma represents 1% of intracranial tumors with a steady increase in prevalence over the past 20 years. The increase in CNS lymphoma is likely secondary to the increased number and longer lifespan of patients with acquired immunodeficiency syndrome (AIDS) and immunosuppression after organ transplantation. Interestingly, the incidence of CNS lymphoma has steadily increased in immunocompetant patients with an increase from 2.5 to 30 cases per 10 million. Deletions of DKN2A are frequently reported in CNS lymphoma.

Macroscopically, primary CNS lymphomas occur within the parenchyma, subependyma, or meninges and can be either circumscribed or irregular. Microscopically, they exhibit diffuse perivascular distribution and infiltrate the walls of blood vessels (perivascular cuffing). The tumor cells are similar in histology to systemic non-Hodgkins lymphoma cells. Primary CNS lymphomas are monoclonal B-cell lymphomas of diffuse large cell or large cell immunoblastic variant. Anti-CD45 antibody staining differentiates CNS lymphoma from other tumor types.

On CT, CNS lymphomas are typically hyper- or isodense to brain with strong contrast enhancement. On MRI, these tumors are usually isointense or hypointense on T1WI, hyperintense on T2WI and display varying degrees of gadalinium enhancement. Low-volume lumbar puncture performed during the workup of CNS lymphoma may reveal high protein, low glucose pleocytosis. While CSF cytology may be diagnostic, stereotactic brain biopsy is often needed for definitive diagnosis.

  1. Choroid Plexus Tumors

The most common tumors of the choroid plexus include choroid plexus papillomas. Choroid plexus carcinomas are rare. Choroid plexus papillomas are most prevalent in patients less then 2 years old and account for less than 1% of all intracranial tumors. Presenting symptoms result from elevated intracranial pressure due to hydrocephalus, and mass effect from tumor growth.

  1. Meningiomas

Meningiomas are typically benign, slow growing extra-axial tumors arising from the arachnoid cap cells of the meninges (Figure 36–14). They can originate wherever arachnoid is present. They are characterized by either their location, or histopathology, and are commonly located along the falx, cortical convexity, and sphenoid bone.

Figure 36–14. Sagittal MRI image demonstrating a contrast enhancing dural-based lesion (arrow) proven to be a meningioma at time of resection.

Menigniomas account for 15%-19% of primary brain tumors and as many as 3% of the population older than 60 years have an intracranial meningioma on autopsy. Their incidence increases with age and peaks by 45 years of age. There is a female-to-male ratio of 2:1.

Inactivation of the NF2 gene found on the long arm of chromosome 22 (22q12.3) is the main genetic event associated with the development of meningiomas. Loss of one copy, truncated, or mutations toward the 5’ end of the NF2 gene occurs in up to 80% of sporadic meningiomas and all patients with neurofibromatosis type 2. Consequently, NF2 patients are more likely to develop intracranial meningioma of varying types. Chromosome 22q contains the NF2 tumor suppressor gene, Merlin.

There are multiple histologic subtypes of meningioma, but they are typically characterized by the presence of densely packed sheets of cells (similar to appearance of normal arachnoid cells), psammoma bodies (whorls of calcium and collagen), intranuclear cytoplasmic pseudoinclusions, and Orphan Annie nuclei (nuclei with central clearing from peripheral migration of chromatin). Radiographically, meningiomas are hyperdense to brain and a broad dural attachment can often be identified. On T2-weighted MRI, most meningiomas are hyperintense and are typically contrast enhancing on both CT and MRI.

  1. Nerve Sheath Tumors and Acoustic Neuromas

Nerve sheath tumors are benign tumors of Schwann cell origin that involve predominantly the fifth, seventh, eighth, and tenth cranial nerves. The most common, vestibular schwannomas (aka acoustic neuromas, AN) originate in the internal auditory canal from the inferior or superior portion of the vestibular nerve at the junction of the central and peripheral myelin. The three most common presenting symptoms include insidious hearing loss, high-pitched tinnitus, and disequilibrium.

AN account for 8%-10% of all intracranial tumors in adults. Most ANs are unilateral, however, patient with neurofibromatosis type 2 commonly have bilateral AN. AN is believed to result from the loss of a tumor suppressor gene located on the long arm of chromosome of 22.

Macroscopically, ANs are lobular, encapsulated, and solid with grayish colored material. Surrounding cranial nerves are often stretched over the capsule of the tumor. Microscopically, these tumors are identical to peripheral schwannomas. They are comprised of Antoni A and Antoni B fibers. Antoni A fibers are dense, narrow, elongated bipolar cells with numerous nuclei and firm cytoplasm. Antoni B fibers are a loose reticulated semi-palisading arrangement of Schwann cells.

CT is useful for distinguishing tumor extension into the bony internal auditory canal. On MRI, ANs are isointense on T1WI without contrast. With gadalinium enhancement they are often homogenously enhancing. The lack of a dural tail, differentiates AN from cerebellopontine angle meningiomas.

  1. Pituitary Tumors

Pituitary adenomas are benign tumors originating from the anterior pituitary gland and represent 10% of intracranial tumors diagnosed. They are most commonly diagnosed between 40 and 50 years of age. They are classified according to their endocrine function or histological staining. Secreting tumors release supraphysiologic levels of hormones that result in distinct clinical syndromes. Hypersecretion of prolactin causes amenorrhea-galactorrhea syndrome in women and impotence in men. Hypersecretion of adrenocorticotropic hormone (ACTH) causes Cushings disease. Hypersecretion of growth hormone causes acromegaly in adults and gigantism in children. Pituitary adenomas can hypersecrete thyrotropin (producing hyperthyroidism) or gonadotropins (leutinizing hormone and follicle stimulating hormone).

Pituitary tumors may exert mass effect on adjacent structures. Optic chiasm compression results in bitemporal hemianopsia. Compression of the pituitary gland itself results in varying degrees of hypopituitarism. Compression upon the cavernous sinus causes ptosis, facial pain, and diplopia from pressure upon cranial nerves III, IV, V1, V2, and VI. Occlusion of the cavernous sinus may cause proptosis and chemosis.

  1. Metastatic Brain Tumors

Metastatic brain tumors originate in from malignancies outside of the CNS that have spread to the brain or spinal cord (Figure 36–15). They are most common brain tumor with a yearly incidence of 100,000-200,000 cases per year in the United States. Autopsy studies show that 20%-25% of patients with cancer have brain metastasis. Metastases occur more frequently in adults in their fifth to seventh decades of life. The most common primary tumors in adults giving rise to CNS metastases are lung, breast, skin, renal, and colon cancers. In children, leukemia and lymphoma, osteogenic sarcoma, and rhabdomyosarcoma are the most common primary tumors that spread to the CNS.

Figure 36–15. Sagittal MRI image demonstrating multiple contrast enhancing lesions (arrows) in the infra- and supratentorial spaces. Biopsy demonstrated metastatic adenocarcinoma.

Histopathology of metastases mirrors that of the primary tumor. MRI is more sensitive than CT in detecting metastases. Metastases are typically seen at the gray-white junction and show varying degrees of contrast enhancement.

Differential Diagnosis

The differential diagnosis of intracranial masses is narrowed through a detailed history and physical exam. Important considerations include patient demographics, chronology of symptoms, past medical history, and specific neurologic deficits. Once imaging is obtained, the list of possible diagnoses can be further refined, as the location of a lesion can suggest its nature. For example, the three most common posterior fossa tumors of childhood include astrocytoma, medulloblastoma, and ependymoma. The most common tumors of the cerebellopontine angle include meningioma, acoustic neuroma, and epidermoid cyst. In addition the pattern of contrast enhancement and whether the lesion appears to arise from within the brain parenchyma or meninges are important considerations in generating an accurate differential diagnosis.

Treatment

  1. Preoperative Medical Management

With few exceptions, surgery is the backbone of the current treatment of brain tumors. The success of surgical intervention depends on adequate preoperative medical management and surgical planning. Steroids are commonly used preoperatively to reduce the symptoms of mass effect and edema caused by the tumor. The timing and dose of steroids varies based on surgeon preference. A common regimen for adults is dexamethasone 6 mg IV or PO every six hours. If mass effect is profound, doses as high as 20 mg every four hours may be considered. Some surgeons believe that it is easier to resect a tumor when peritumoral edema is minimized by preoperative decadron administration.

The use of anticonvulsants in brain tumor patients at presentation, preoperatively, and postoperatively is somewhat controversial. Without question, patients presenting with seizures attributed to a brain tumor should be initiated on an anticonvulsant. However, with few exceptions, there is no data to suggest prophylactic use of anticonvulsants reduces the risk of new onset seizures in brain tumor patients. Among the exceptions are: (1) tumor involvement in highly epileptogenic areas such as the motor cortex, (2) low-grade gliomas, which carry a high risk of seizures, (3) patients with metastatic lesions that commonly invade the cortex, and (4) patients with both metastases and leptomeningeal spread.

Because of a favorable toxicity profile and cost, phenytoin is the first line antiepileptic agent. Phenytoin may cause GI upset and should be administered with a H2 blocker or proton-pump inhibitor. Phenytoin levels must be monitored to ensure a therapeutic serum drug concentration. Leviteracetam is an alternative used for patients if there is potential for drug interactions related to induction of the P450 system by phenytoin. In contrast to phenytoin, levels of leviteracetam need not be monitored.

  1. Surgical Considerations

The surgeon must decide whether the goal of intervention is obtaining biopsy only, subtotal resection, or attempting gross total resection. With few exceptions, gross total resection offers the best chance of survival and is the preferred treatment. Tumor resection requires careful consideration of a number of key factors including: (1) tumor size, (2) location, (3) gross, radiographic, and pathologic characteristics, (4) sensitivity to radiation, and importantly, (5) the medical and neurologic status of the patient.

The timing of surgery is important in preoperative planning. Patients who present with rapid deterioration due to elevated intracranial pressure typically require prompt intervention. Tumor growth can be brisk in patients with large, high-grade tumors and small increases in tumor volume can cause a profound increase in intracranial pressure. Rapid intervention may also be necessary in the setting of obstructive hydrocephalus. Cerebrospinal fluid diversion (typically via ventriculostomy) is an alternative to urgent tumor resection in patients with obstructive hydrocephalus secondary to tumor growth. In cases where tumor burden is not causing profound neurologic deficit or elevated intracranial pressure, resection can be arranged as a semi-elective procedure.

Once in the operating room for brain tumor resection, a number of important principles of positioning are vital to a successful resection. Most tumor resections require immobilization of patient’s head in a Mayfield head holder. The position should be selected create the most direct access to the lesion while avoiding risk to other bodily structures. Positioning should promote venous drainage from the lesion and the cranial compartment by ensuring the jugular veins are not compressed and that the head is elevated. In most cases, surgeons prefer to position the patient with the operative corridor perpendicular to the floor. This approach generally minimizes brain retraction and is the most ergonomic for the surgeon. Pressure points of the remainder of the patient should be padded especially thoroughly given the lengthy nature of some tumor resections.

The shape of the skin incision and bone flap is dependent on the desired approach, size of the lesion, and surgeon preference. Small tumors can be adequately exposed and resected via linear or curvilnear incisions with a small bone flap. Resection of deep lesions, especially those involving the skull base, often require creation of a sizable scalp flap and removal of a large window of bone. Whenever possible, incisions should be planned behind the hairline, minimizing the amount of hair removal in deference to cosmetic concerns. The placement of the incision is largely determined by the location of the lesion. Frameless stereotaxy, a technology that relies on a three-dimensional rendering of the patient’s preoperative MRI, can enable accurate tumor localization based on preoperative imaging and can be helpful in minimizing the size of the incision. Standard approaches to intracranial lesions minimize the morbidity of exposure by limiting the risk to key neural and vascular structures.

The central goal of brain tumor surgery is maximizing the removal of neoplastic tissue while minimizing collateral damage to surrounding normal brain and vascular structures. Standards for achieving this goal vary based on tumor type. For example, the goal of the resection of a high-grade glioma is to remove all enhancing portions of the tumor, in contrast to the goal for the resection of a low-grade glioma to remove the tissue that appears abnormal on T2-weighted MRI. Several large retrospective studies published in the last decade suggest that duration of survival is directly related to extent of resection or the proportion of tumor removed at the time of surgery. The goal for resection of meningioma is to remove both the tumor and its dural origin. Metastatic tumors are typically well demarcated and often encapsulated and the goal is to remove the entire tumor.

One of the central challenges in brain tumor surgery is that neoplastic tissue that is easily detected on MRI is often virtually indistinguishable from normal brain. Several studies evaluating the extent of brain tumor resection highlight the fact that in many cases, especially in diffusely invasive brain tumors, a significant amount of residual tumor remains even after gross total resection. Moreover, surgeons have a limited ability to predict when all resectable tumor has indeed been removed. Consequently, a variety of technologies have been developed to improve surgical outcomes. Stereotactic navigation is utilized to improve extent of resection but there is little evidence that it can improve extent of resection. Retrospective analyses suggest that intraoperative MRI, an approach in which brain tumor resection is performed in a highly specialized surgical suite containing an MRI machine, improves extent of resection. Fluorescent and visible dyes have been proposed as a means of identifying tumor margins intraoperatively for more than 60 years. Recently a phase III clinical trial has demonstrated that the fluorescent dye 5-ALA may improve the extent of resection and six-month progression-free survival in glioblastoma patients. A number of efforts are underway to use dye-based and label-free intraoperative microscopy to improve the surgeon’s ability to distinguish tumor-infiltrated brain from noninfiltrated tissue.

Intraoperative electrophysiologic monitoring of brain activity is often used in tumors within eloquent cortex to determine a safe route for exposure of tumor and the safe limits for extent of resection. Electrophysiologic motor mapping can be performed with the patient under general anesthesia or awake. In asleep motor mapping, specific regions of the motor cortex or corticospinal tract are stimulated with electrical current while the electromyographic response in target muscles is recorded. In awake motor mapping direct electrical stimulation of motor cortex or corticospinal tracts is performed while a patient is asked to perform specific tasks. An arrest of motor activity with direct electrical stimulation suggests that the portion of brain being stimulated is involved in motor function. Similarly, in awake language mapping specific regions of the brain, commonly the dominant frontal and temporal lobes are stimulated to look for speech arrest. Awake language mapping of brain tumor patients has broadened our understanding of the organization of the human language cortex and circuits.

Following resection, meticulous attention is paid to achieving hemostasis in the operative corridor to minimize the risk of postoperative hemorrhage. Whenever possible, to diminish the risk of cerebrospinal fluid leak, a watertight dural closure is performed. The bone flap is replaced and the galea is reapproximated. Scalp closure that omits closure of the galea provides little strength and raises the risk of dehiscence.

  1. Postoperative Management

Following resection patients are observed closely in an ICU setting, typically for overnight, where serial neurological examinations are carried out. Depending on the extent of resection, steroids may be tapered over the days following surgery. Anticonvulsants are continued in patients who have a history of seizures and, when there has been extensive brain dissection, they may be continued for 1-4 weeks following surgery. Given the prognostic significance of extent of resection for glioma patients and the difficulty in detecting residual tumor during surgery, it is becoming standard practice for surgeons to obtain postoperative MRI imaging with contrast to evaluate for residual tumor within 24 hours of resection. When there is a low suspicion of residual tumor or when further surgery is not possible, postoperative imaging may be deferred.

  1. Adjuvant Therapies

Surgical resection is cornerstone of brain tumor therapy but it is rarely capable of eradicating all tumor cells. Furthermore, resection may not be favored when eloquent structures are likely to be damaged. Adjuvant radiation and chemotherapy regimens have been developed to address the inability of current surgical techniques to reliably eradicate residual or unresectable tumor.

  1. Radiation—Radiation kills tumor cells by directly damaging cellular structures, inducing lethal mutations in cellular DNA and by activating pathways for programmed cell death. Radiation can be delivered to brain tumors in a fractionated manner, which allows normal tissue repair between treatments and increases the toxicity of the radiation to tumor tissue.

Regimens for radiation therapy of brain tumors vary with tumor type. The optimal dose and timing of adjuvant radiation therapy for low-grade glioma is controversial. Typically, radiation is reserved in low-grade glioma until there is evidence of tumor progression or neurologic deterioration. Early radiation therapy is suggested in elderly patients, for tumors that have crossed the midline and in the setting of intractable seizures. Because of the cognitive consequences of radiation therapy, it may be delayed in younger patients until there is suspicion of recurrence

For high-grade glioma, the findings of a study conducted by the Brain Tumor Cooperative Group (BTCG) demonstrated an increase in survival in HGG patient undergoing radiation therapy plus surgery compared to surgery alone from 14 to 31 weeks. A landmark study on the use of radiation in conjunction with temozolomide has defined the current standard for radiation therapy in glioblastoma patients: 2 Gy given 5 days per week for 6 weeks, totaling 60Gy.

Two important clinical trials have established the standard therapy for metastatic lesions. Currently, acceptable care of patients with brain metastases consists of resection followed by whole brain radiation therapy, or stereotactic radiosurgery (SRS), in which a high dose of radiation is administered to the tumor bed, in addition to whole brain radiation therapy. The use of radiation therapy has been extensively investigated. Currently, radiation therapy is indicated as an adjunct to surgery in the setting of recurrent meningioma or subtotal resection. Occasionally in a poor surgical candidate or if a meningioma is in a location carrying high surgical risk, radiation therapy may be used in isolation.

  1. Chemotherapy—Clinical trials of chemotherapy for low-grade glioma have been limited and the use of chemotherapy for this remains experimental. The one exception is low-grade gliomas, typically those with 1p deletion and oligodendroglial lineage, which are particularly sensitive to PCV (procarbizine, carmustine, and vincristine) or temozolomide. In contrast, a recent clinical trial in patients with proven GBM comparing radiation alone to radiation in combination with the oral alkylating agent temozolomide demonstrated a modest but significant increase in survival from 12.1 to 14.6 months. The current standard of care for patients with GBM combines radiation therapy with oral temozolomide. Chemotherapy has not shown any benefit in the treatment of brain metastases and meningiomas. Future efforts in the development of novel chemotherapeutic agents are focused on developing novel inhibitors of signaling pathways that are active only in brain tumor cells. Developing novel methods for the delivery of traditional chemotherapeutic agents is also an area of active research.

Complications

Patients with primary and metastatic brain tumors are at risk of developing postoperative medical as well as surgical complications. Sawaya proposed the most common classification scheme for complications associated with brain tumor surgery in 1998. In a case series of 400 craniotomies for treatment of brain tumors, complications were classified as neurological, regional, and systemic. Neurologic complications are outcomes that produce visual field, motor, sensory, or language deficits. Neurologic complications are the result of injury to normal brain structures, cerebral edema, hematoma, or vascular injury. In most series, the risk of a new neurologic deficit after craniotomy for resection of an intrinsic brain tumor ranges from 10% to 25%. The risk factors for adverse neurologic outcomes include older age (greater than 60 years), deep tumor location, tumor proximity to eloquent regions, and low functional performance score (Karnofsky score less than 60%). Neurologic complications can be minimized by individualizing the surgical approach for each patient, cortical mapping techniques, minimizing excessive brain retraction, meticulous hemostasis, and early identification of major venous structures.

Regional complications are those related to the surgical wound or brain parenchyma, without neurologic deficit. They occur in 1%-5% of patients undergoing craniotomy for resection of an intrinsic brain tumor. Regional complications include wound infections, pneumocephalus, Cerebrospinal fluid fistula, hydrocephalus, seizure, brain abscess/cerebritis, meningitis, and pseudomeningocele. These complications occur more readily in the elderly. Posterior fossa location and reoperations are associated with a higher rate of pseudomeningocele, CSF fistula, hydrocephalus, and wound infections. Postoperative wound infections and cellulitis occur in 1%-2% of patients after supratentorial craniotomy. They typically result from skin bacterial contamination (Staphylococcus aureus and Staphylococcusepidermidis). The risk of postoperative seizures following supratentorial craniotomy is 0.5%-5%. Prophylactic antiepileptic drugs can be routinely used in the postoperative period; however, their dose and duration is an area of controversy.

Systemic complications include all generalized adverse events, including deep vein thrombosis (DVT), pulmonary embolus, pneumonia, urinary tract infections, myocardial infarction, and sepsis. These medical complications occur in 5%-10% of patients undergoing craniotomy for removal of an intrinsic brain tumor and are more prevalent in older patients (greater than 60 years) and neurologically impaired patients (Karnofsky score less than 60%). DVT is the most common complication occurring in 1%-10% of patients within the first month after a craniotomy. Patients with systemic cancer, glioblastoma multiforme, meningiomas, lower extremity paralysis, bed rest, and prolonged surgery are at particularly increased risk of developing a DVT or pulmonary embolus. Early postoperative mobilization, intermittent compression devices, and postoperative anticoagulation with low-molecular-weight heparin have decreased the incidence of postoperative DVT.

Craniotomy for resection of brain tumor can be performed safely and most complications can be prevented with careful preoperative planning, meticulous surgical technique, and attentive postoperative care.

Prognosis

The prognosis of brain tumor patients varies based on a number of factors, including, but not limited to, general functional status at the time of diagnosis, tumor type, location, and age.

  1. Glioma

The prognosis of glioma patients is determined by tumor grade, age, extent of resection, Karnofsky performance status, and treatment response. Improvement in survival when radiographically complete resection is achieved is greatest for those with low-grade lesions. While more modest, high-grade glioma patients with radiographically complete resection also have survival improvement compared to incomplete resection. Achieving gross total resection improves survival by lowering the risk of recurrence and reducing tumor cell burden to levels that can be eradicated or controlled with adjuvant therapy.

  1. Meningioma

In general, the prognosis of meningioma patients is more favorable than that of glioma patients. The prognosis of meningioma patients is determined by the extent of resection and tumor grade. The Simpson classification system stratifies meningioma patients into outcome groups based on extent of resection. Patient age, extent of surrounding structure invasion, male gender, genetic factors, and tumor grade are among the factors that are linked to prognosis. Recurrence has been estimated to occur in approximately 20% of patients with benign meningiomas but is much more common in higher grade lesions.

Metastases

The survival of patients with untreated brain metastasis is quite poor (1-2 months), but survival can be prolonged by 4 or more months with optimal surgical and radiation therapies. Extent of extracranial disease is a key prognostic factor in patients with brain metastases. Age and Karnofsky performance status have an important bearing on overall survival.

Kool M, Korshunov A, Remke M, et al: Molecular subgroups of medulloblastoma: an international meta-analysis of transptome, genetic aberrations, and clinical data of WNT, SHH, Group 3, nad Group 4 medulloblastomas. Acta Neuropathol 2012;123:473-484.

Prabhu VC, Khaldi A, Barton KP, Melian E, Schneck MJ, Primeau MJ, Lee JM: Management of diffuse low-grade cerebral gliomas. Neurol Clin 2010 Nov;28(4):1037-1959.

Taylor MD, et al: Molecular subgroups of medulloblastoma: the current consensus. Acta Neuropathol 2012;123:465-472.

TUMORS OF THE SPINE AND SPINAL CORD

Anthony C. Wang, MD Khoi D. Than, MD Paul Park, MD

INTRODUCTION

Neoplastic pathology affecting the spinal cord is uncommon in the general population but it is an important consideration in the evaluation of patients presenting with neck and/or back pain with or without associated radicular symptoms, sensorimotor deficits, and bowel or bladder dysfunction. An estimated 15% of primary CNS tumors are intraspinal, and most of these are benign. Since the first resection of a spinal cord tumor was reported in 1888, surgery has remained a mainstay of treatment in the majority of spinal tumors, though radiotherapy and chemotherapy have demonstrated benefit in a growing number of instances.

Spinal tumors are differentiated based upon their locations relative to three anatomic compartments. Extradural tumors are located outside of the thecal sac, and arise either from the osseous spine or epidural space. Intradural-extramedullary tumors occur within the thecal sac but are outside of the neural tissues of the spinal cord. These tumors most commonly develop from the leptomeninges or nerve roots. Intramedullary tumors are found within the spinal cord and originate from either the spinal cord parenchyma or pia mater.

In addition to gender and age of presentation, localization of the lesion to the cervical, thoracic, lumbar, or sacrococcygeal spine aids in refining the differential diagnosis as certain tumors demonstrate a predilection for particular regions of the spinal column.

Clinical Presentation

Symptoms experienced by patients presenting with spinal tumors are more commonly produced by compression than direct invasion of the spinal cord or nerve roots. Classically, the pain associated with neoplasms is unremitting, worse in the supine position, and more noticeable at rest or in bed. Hence, the patient may wake up at night due to pain. Although the progression of symptoms can be insidious, radicular pain, motor weakness, paresthesias, or anesthesia can frequently occur as a result of nerve compression. Long tract myelopathic findings such as ataxia, hyperreflexia, spasticity, fasciculations, sensorimotor loss, or sphincter dysfunction can be caused by spinal cord compression. Recognized unusual findings include muscle wasting or hyporeflexia, referred pain, autonomic changes such as in Horner’s syndrome, and Brown–Séquard hemicord syndrome. Fracture and deformity causing axial pain often occur as well, and may be the presenting complaint.

Extradural tumors frequently involve the osseous spine, and so most typically present with axial pain, often increased by motion or Valsalva maneuver. Signs and symptoms of neural compression occur secondarily. Intradural-extramedullary tumors present most commonly with motor deficits and other long tract disturbances. Intramedullary tumors can demonstrate an insidious succession of symptoms including neuralgic pain that can progress to myelopathy.

Diagnosis

  1. Radiographic Evaluation

MRI is currently the primary mode of assessment of spinal tumors (Figure 36–16). Contrast-enhanced T1-weighted sequences are the standard initial imaging modality for the evaluation of suspected extradural, intradural-extramedullary, and intramedullary lesions. Nearly all intramedullary lesions demonstrate contrast uptake, and the resolution that MRI provides is typically sufficient for determination of margins and infiltration. The addition of angiography is indicated for the suspicion of a vascular pathology.

Figure 36–16. Sagittal T1-weighted image with contrast of the upper thoracic spinal cord demonstrating an intradural, intramedullary lesion found on biopsy to be a high-grade glioma.

If MRI is not obtainable, then more invasive methods are used. Myelography, which was once the modality of choice for imaging the spinal canal, provides excellent structural detail. Fusiform cord widening, a dumbbell-shaped deformity, or complete blockage are classic findings suggesting the presence of a mass lesion. Plain, contrast-enhanced, and postmyelography CT imaging can aid in the assessment of bony spine anatomy. Nuclear scintigraphy is used primarily in the identification of skeletal metastases. Catheter angiography can be employed in the evaluation of suspected vascular lesions. Plain x-rays of the spine have limited utility in the initial assessment of spinal tumors. Findings such as enlarged intervertebral foramina and interpedicular spaces, or bone erosion with scalloped edges suggest the presence of an enlarging mass.

  1. Laboratory Analysis

Lumbar puncture can provide additional clues to the presence of a spinal cord tumor. Elevated CSF protein is present in approximately 95% of cases, though CSF glucose is normal. Xanthochromia and the presence of fibrinogen causing clotting can also occur. Specific CSF and serum immunocytochemical tests can aid the diagnosis of specific neoplasms.

  1. Differential Diagnosis

Among all spinal tumors, extradural lesions are discovered with the greatest frequency, comprising an estimated 55%-60%. A vast majority of these are metastatic lesions via hematogenous dissemination along Batson’s venous plexus. The most common primary sources of vertebral metastases are lung, breast, and prostate.

Benign spine tumors of the osseous spine are very rare, and may be incidental findings, or cause pain, radiculopathy, myelopathy, spinal instability, or deformity. Biopsy is utilized at times, and appropriate management may be observational, as is common in the case of eosinophilic granuloma, ablative, as can be performed for osteoid osteoma, or surgical. Primary benign spine tumors can be classified using the Enneking system.

Approximately 5% of bone malignancies involve the spine, the four most common being osteosarcoma, chondrosarcoma, Ewing’s sarcoma, and chordoma. Survival varies significantly by pathology. The high rates of recurrence, limited survival duration, and functional morbidity associated with these tumors support the aggressive multimodal strategies typically employed in treatment of these tumors. In some cases, complete resection with negative margins, or even en bloc resection, is possible. In other cases, surgical resection may serve to decrease tumor burden, thereby improving the efficacy of adjuvant treatments such as chemotherapy and radiotherapy. At times, surgical intervention is necessary for functional indications, either for neural decompression or spinal stabilization.

Only 0.5% of tumors involving the spinal column are primary neoplasms. Of the intradural tumors, an estimated 70% are extramedullary. Schwannoma, neurofibroma, and meningioma comprise the bulk of these cases. The nerve sheath tumors, schwannoma and neurofibroma, are usually benign, and can demonstrate a typical dumbbell-shaped appearance, caused by the neuroforamina through which they sometimes pass. They are an important consideration when deviation of the pleural reflection on x-ray suggests a posterior mediastinal mass. Schwannomas are well-differentiated and typically grossly resectable; however, neurofibromas cannot be resected completely from its parent nerve. In the setting of type-1 neurofibromatosis, suspicion of multiple neurofibromas, meningioma, and ependymoma should be heightened. Schwannoma, neurofibroma, and meningioma multiplicity is associated with type-2 neurofibromatosis. Meningiomas are seen most often in the thoracic spines of middle-aged women, and arise from persistent arachnoid cap cells.

Intramedullary tumors comprise only approximately 10% of all spinal tumors, and arise most commonly in the cervical segment of the spinal cord. Glial tumors are the most common of these, with ependymomas occurring twice as frequently as astrocytomas in adults. In children, the relationship is reversed, with astrocytomas being twice as common as ependymomas. Myxopapillary ependymomas form at the conus medullaris and filum terminale, and are a very common tumor to be found at this location.

Treatment

Treatment options are tailored to the patient. Some patients present with axial pain absent of radiculopathy while others present with mild or stable evidence of spinal cord compression causing myelopathy. Still others present with rapidly progressive course of neurologic deterioration.

Generally, primary tumors of the spinal column are best treated by complete resection, if possible. In the vertebral column, en bloc resection is frequently the goal of surgery, whereas with any involvement of the spinal cord, gross total resection is the desired treatment. Symptomatic deformity or spinal instability warrants consideration for surgical fixation, as functional deterioration profoundly affects survival.

Extradural metastases are the most frequently seen tumors of the spinal column, and their treatment is complicated by many factors. Decompression of the spinal cord is the initial surgical consideration. Resection followed by adjuvant radiation has shown promising oncologic results. Spine stability must then be addressed depending on the patient’s overall prognosis.

In 2007, the WHO updated a comprehensive classification of neoplasms affecting the CNS based upon the specific cell type from which each tumor arises, which guides therapy and prognosis in cases of most intramedullary tumors. Overall, functional outcome depends heavily on the severity and duration of neurologic symptoms at the time of presentation, while survival outcome depends largely on the pathology in question, and less so, on the ability to gain adequate surgical resection.

Intradural-extramedullary tumors are generally best treated by surgical resection; outcomes appear to depend heavily upon extent of resection, though preservation of preexisting neurologic function remains the primary objective in surgery. Determination of a dissection plane between tumor and spinal cord is the initial aim of any resection procedure. Overall, prognosis is excellent in nearly all cases of spinal intradural-extramedullary neoplasms, unless paraplegia is the presenting condition. In contrast, recurrence of infiltrative neoplasms is very common, and progression to paralysis is common. Progressive neurologic deterioration demands particular focus on facilitating appropriate surgical therapy.

Binning M, Klimo P Jr, Gluf W, Goumnerova L: Spinal tumors in children. Neurosurg Clin N Am 2007;18(4):631-658.

Louis DN, Ohgaki H, Wiestler OD: The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 2007;114(2):97-109.

Mukherjee D, Chaichana KL, Parker SL, Gokaslan ZL, McGirt MJ: Association of surgical resection and survival in patients with malignant primary osseous spinal neoplasms from the Surveillance, Epidemiology, and End Results (SEER) database. Eur Spine J 2013 Jun;22(6):1375-82.

Ozawa H, Kokubun S, Aizawa T, Hoshikawa T, Kawahara C: Spinal dumbbell tumors: an analysis of a series of 118 cases. J Neurosurg Spine 2007;7(6):587-593.

Thakur NA, Daniels AH, Schiller J, Valdes MA: Benign tumors of the spine. J Am Acad Orthop Surg 2012;20(11):715-724.

Zadnik PL, Gokaslan ZL, Burger PC, Bettegowda C: Spinal cord tumours: advances in genetics and their implications for treatment. Nat Rev Neurol 2013 May;9(5):257-266.

PITUITARY TUMORS

Wajd N. Al-Holou, John E. Ziewacz, Ariel Barkan, William F. Chandler, Stephen E. Sullivan

Clinical Considerations

The pituitary gland, involved in the regulation of the major hormonal axes of the body, is located in the sella turcica (“Turkish Saddle”) of the sphenoid bone and is comprised of the anterior pituitary (adenohypophysis), posterior pituitary (neurohypophysis), and functionally insignificant pars intermedia separating the two lobes. The anterior pituitary secretes Prolactin, ACTH, TSH, LH, FSH, and GH. The posterior lobe is a repository of the hypothalamic hormones Oxytocin and Anti-Diuretic hormone (ADH, vasopressin).

Pituitary tumors are typically benign adenomas that arise from the anterior lobe of the pituitary gland. They comprise approximately 10% of CNS tumors. With the increasing use of brain MRI, there has been an increase in the diagnosis of incidentally discovered pituitary adenomas, known as incidentalomas, with nearly a 23% imaging prevalence. Pituitary adenomas are classified by both their secretory status (secreting vs. nonsecreting), and their size. Microadenomas have diameters up to 1 cm, while tumors greater than 1 cm are termed macroadenomas (Figure 36–17). Secreting (endocrine-active) tumors are further classified by which hormone is being hypersecreted and the resulting clinical syndrome.

Figure 36–17. Magnetic resonance imaging A. sagittal plain and B. coronal plain of a pituitary macroadenoma.

Clinical Findings

Clinical findings in patients with pituitary adenomas are caused by compression of surrounding structures, decrease in pituitary function, hypersecretion of pituitary hormones, and rarely, acute pituitary failure (apoplexy). Pituitary tumors can also be asymptomatic, and discovered incidentally.

  1. Compression

Compression caused by slowly expanding adenomas causes a variety of findings which include headache, bitemporal hemianopsia (decreased perception of the lateral visual fields), diplopia, and hypopituitarism (decreased secretion of pituitary hormone).

Bitemporal hemianopsia results from the upward extension of the tumor with associated compression of the decussating nasal fibers of the optic chiasm, which lie directly above the pituitary gland. This can be detected on examination as red color desaturation in the temporal visual fields. Without treatment, visual findings can progress to loss of visual acuity and eventual blindness.

Hypopituitarism results from compression and injury of the normal pituitary cells by the expanding mass. Diminished secretion of GH, LH, and FSH occurs early, while diminished secretion of TSH and ACTH occurs later.

Symptoms related to decreased production of LH and FSH manifest as hypogonadism: loss of libido and erections in men and amenorrhea in women. In men, this is often overlooked and only recognized in retrospect. The decreased production of LH and FSH can be due to either direct compression of the gland, or to compression of the stalk, which releases the tonic inhibition of prolactin secretion by dopamine from the hypothalamus. This results in a mild increase in prolactin (usually less than 150 ng/mL) which inhibits gonadotropin release. This can cause galactorrhea as well, and is the mechanism of the amenorrhea/galactorrhea syndrome.

Decreased production of TSH causes classic findings of hypothyroidism, including cold intolerance, weight gain, fatigue, coarse hair, and myxedema. Decreased production of ACTH causes hypocortisolism resulting in fatigue, slow return to health after minor illness and orthostatic hypotension. As the body’s intrinsic stress response is diminished with the loss of ACTH production, in rare circumstances, this can result in cardiovascular collapse when the individual is under extreme stress.

  1. Hypersecretion

Functional adenomas produce symptoms related to hypersecretion of a specific hormone. They can also produce findings related to compression, but this is less common than in nonfunctional tumors, given that they are often discovered at an earlier stage because of the clinical findings caused by hypersecretion.

Prolactinomas are the most common functional pituitary tumor, accounting for nearly 30%-40% of pituitary tumors. Prolactinomas cause hypersecretion of prolactin. Interruption of the pituitary stalk can cause hyperprolactinemia as well, but levels are typically more modest (150 ng/mL or less). Levels greater than 300 ng/mL are virtually always associated with prolactin secreting tumors. Symptoms of hyperprolactinemia are gender-dependent and may include galactorrhea, amenorrhea (via suppression of gonadotropins), diminished libido, and infertility.

Growth hormone-secreting tumors are the next most common endocrine active tumors. They cause gigantism in children and acromegaly in adults. Acromegaly is characterized by classic physical changes such as soft tissue and skeletal overgrowth, prognathism, widely spaced teeth, macroglossia, as well as nerve entrapment syndromes and arthropathy. It also results in conditions associated with increased morbidity and mortality including cardiomyopathy, sleep apnea, and glucose intolerance. Acromegaly is insidious in onset and is often not noticed by the patient’s friends or family. Patients often note a change in shoe size as an adult, an inability to wear a wedding ring, or recognize a striking physical change when comparing a recent to a past photograph.

Tumors that secrete ACTH are the cause of Cushing disease, which refers specifically to hypercortisolism caused by a functional, ACTH-producing, pituitary tumor. The resultant signs and symptoms of hypercortisolism from any etiology are termed Cushing syndrome. Characteristic findings include weight gain (with centripetal fat distribution), “buffalo hump,” that is, redistribution of fat to the posterior neck, purple abdominal striae, thin skin with easy bruising, round “moon” face, glucose intolerance, hypertension, osteoporosis, poor wound healing, psychiatric disturbance (depression, mood lability), amenorrhea, impotence, and hyperpigmentation (only with elevated ACTH). These patients have a significant increase in morbidity and mortality if left untreated, with reports of a fourfold to fivefold increase in mortality in some studies. Gonadotropin producing tumors are rare and result in the hypersecretion of LH and FSH. They are often clinically silent, and usually behave like nonfunctioning tumors. In women, they can produce amenorrhea and infertility.

TSH-producing tumors are very rare, accounting for less than 1% of secretory tumors. Clinical findings are those of classic hyperthyroidism including weight loss, tachycardia, heat intolerance, anxiety, and tremor. As opposed to classical thyrotoxicosis (multinodular goiter, Graves’ disease, and iatrogenic causes) in which plasma TSH concentrations are undetectably low, in patients with thyrotropinomas TSH levels are “normal” or frankly elevated despite high free T4 and T3 concentrations.

  1. Acute Pituitary Failure (Apoplexy)

Acute pituitary failure is usually the result of hemorrhage and/or necrosis in a preexisting adenoma. Symptoms are abrupt and may result in severe headache, visual disturbance, ophthalmoplegia, and change in mental status. Pituitary apoplexy is often considered a neurosurgical emergency, especially in the setting of visual loss. Urgent decompression is warranted for acute or continuing neurologic decline, in order to prevent untoward neurologic consequences, especially blindness. Pituitary hormonal failure almost always accompanies an apoplectic event, and rapid administration of corticosteroids is necessary to avoid cardiovascular collapse.

  1. Stalk Compression

Symptoms related to stalk compression are rare in the setting of pituitary adenomas and consist of prolactin elevation and loss of antidiuretic hormone (ADH). Diabetes insipidus results from impairment of hypothalamic input and the resultant diminished release of ADH. This causes an inability to concentrate the urine and resulting dehydration and hypernatremia. Symptoms include frequent high-volume urination and excessive thirst. This can be very dangerous in the setting of a patient with an impaired thirst mechanism. Stalk compression is more common with tumors directly involving the stalk such as craniopharyngioma and Langerhans cell histiocytosis.

Mild prolactin elevation results from blocking hypothalamic dopamine from accessing the pituitary prolactin-producing cells. This release of tonic inhibition results in a modest elevation of prolactin (usually less than 150 ng/mL). This is called the “stalk effect” and may result in amenorrhea and galactorrhea.

Differential Diagnosis

The differential diagnosis for sellar and parasellar masses is broad. The hallmarks of distinguishing among the multiple possibilities are a detailed history and physical examination, MRI scan, and pituitary hormone testing. The history and physical examination should be aimed at symptoms related to pituitary hyper- or hyposecretion, and compression of surrounding structures. A detailed visual field examination should be included in the evaluation of any potential pituitary mass. Often, formal visual field testing by an Ophthalmologist is warranted. MRI should be obtained with and without contrast, with thin cuts through the sellar region in coronal and sagittal planes. This will identify the size, configuration, and extent of invasion of a pituitary tumor. Laboratory testing should include evaluation of anterior pituitary hormones, directly or indirectly, in order to establish a hyper- or hyposecretory state. These include prolactin, 8:00 AM cortisol, free T4, TSH, IGF-1, GH, LH, FSH, and testosterone (in men). If symptoms of diabetes insipidus are present, serum sodium and osmolality, and urine osmolality should be obtained to confirm the diagnosis. An overnight water deprivation test may be warranted to evaluate if the patient is able to concentrate urine appropriately.

If a hypersecretory state is not detected on detailed laboratory testing, the lesion is likely a nonfunctional pituitary adenoma or other entity. Other possibilities include craniopharyngioma, Rathke’s cleft cyst, or meningioma. A number of other less common lesions are in the differential including metastasis, epidermoid, dermoid, germinoma, abscess, Langerhans histiocytosis, choristoma, chondrosarcoma and chordoma. Rarely, a giant aneurysm can mimic a pituitary tumor.

If a hypersecretory state is discovered, the diagnosis depends on the hormone that is being hypersecreted. It is important to distinguish a primary pituitary lesion causing hypersecretion from a lesion in another location, such as an ACTH-secreting carcinoid lung tumor.

Prolactin levels greater than 200 ng/mL indicate the presence of a prolactinoma. Modest elevations in prolactin, less than 150 ng/mL, can be caused by compression of the pituitary stalk, antidopaminergic drugs, estrogen excess (usually from oral contraceptives), chest wall lesions, primary hypothyroidism, and hypothalamic damage. In evaluating prolactin in patients with giant adenomas (> 3.5 cm), it is important to include serial dilutions up to 1:1000 in addition to the undiluted serum value. The purpose of this is to obviate a false negative value in cases of extremely elevated prolactin levels caused by the so-called “hook effect.” This results from excess prolactin binding to both antibodies of the laboratory assay, causing a lack of formation of the complexes identified in the current assays. If the diluted level is still elevated, one can conclude there is a truly elevated prolactin level.

If there is concern for hypercortisolism, a 24-hour urinary free cortisol should be obtained to diagnose hypercortisolism. This should always be repeated for confirmation of the abnormal elevation. If this value is clearly elevated then hypercortisolism is definite and search is initiated to find the etiology of the hypercortisolism. If the level is equivocal and suspicion of hypercortisolism is still strong, then a low- and high-dose dexamethasone suppression test may be employed.

Once hypercortisolism is confirmed, the location of the pathology must be identified. Identification of an adenoma on MRI will obviate the need for further endocrinological studies. Endogenous hypercortisolism is either ACTH-dependent, most commonly from a pituitary adenoma or ectopic ACTH from a pulmonary malignancy, or ACTH-independent, resulting from an adrenal source (adrenal adenoma or carcinoma). Tests useful in confirming an ACTH-dependent tumor include a 4 PM erum ACTH. If ACTH level is elevated, then the source needs to be identified. The low- and high-dose dexamethasone suppression test (Liddle test) can also help to confirm a pituitary etiology. In Cushing disease serum cortisol is not suppressed with the low dose of 0.5-mg dexamethasone q 6 hours for 2 days, but does suppress with the higher dose of 2.0 mg q 6 hours for 2 days. In patients with ectopic sources of ACTH production, cortisol level is not suppressed even with the high-dose dexamethasone suppression test. In cases of proven ACTH-dependent hypercortisolism in which a pituitary tumor is not seen on MRI, inferior petrosal sinus sampling (IPSS) is required to confirm a pituitary etiology, as well as to help localize the side of the tumor. IPSS can identify a pituitary source of hypercortisolism in 98%, but it is not very effective in identification of laterality, being correct in no more than 69% of cases.

Laboratory diagnosis of GH secreting tumors is often indirect. This is because the secretion of GH is pulsatile and levels in patients without a hypersecreting tumor can often be elevated at certain times during the day. Patients with active acromegaly may have completely normal random GH levels. IGF-1 (somatomedin-C) is the gold standard for diagnosing GH secreting tumors. IGF-1 is produced in the liver and other organs and is dependent on GH for its production. Serum levels of IGF-1 are relatively stable and provide a better confirmatory test than the more volatile GH.

LH and FSH secreting tumors are diagnosed by serum elevated levels in the setting of a pituitary mass diagnosed on MRI.

TSH secreting tumors are rare and must be distinguished from other pathologic entities on the hypothalamic-pituitary-thyroid axis. Secondary hyperthyroidism (that produced by a pituitary tumor) is diagnosed by an elevated TSH level as well as an elevated free T4 level. In primary hyperthyroidism, free T4 would be elevated and TSH would be suppressed via a negative feedback mechanism.

Treatment

Treatment of pituitary tumors depends on their size, hormonal characteristics, and level of invasiveness. Micro- or macroadenomas that are nonsecreting and are not causing symptoms of compression may be treated conservatively, and followed with serial imaging. The natural history of nonsecreting microadenomas is often benign.

Nonsecreting macroadenomas that cause signs of compression are currently treated surgically. The typical surgical approach is a transsphenoidal route. This is generally performed transnasally, using the operating microscope and/or the endoscope. Recent advances in frameless stereotactic guidance and endoscopic techniques have allowed a greater ability to localize lesions intraoperatively and provide a more minimally invasive approach to treatment of pituitary adenomas. The use of endonasal endoscopic approaches has increased in recent years and provides wider intraoperative visualization, but may be associated with increased intranasal morbidity. Nonetheless, both endoscopic and microscopic endonasal approaches are effective surgical treatment modalities, and the choice of technique is often surgeon-dependent.

Occasionally, if an adenoma has an unusual amount of suprasellar extension, or lateral extension a craniotomy may be necessary. In some cases, lesions that invade the cavernous sinus or have a significant amount of extrasellar extension are not surgically accessible by any approach and radiation therapy is often necessary to provide tumor control.

Patients with hypopituitarism due to compression who do not experience return of pituitary function postoperatively require hormone replacement on a chronic basis. If chronic steroid replacement is necessary, it is important that patients be provided with a medic-alert bracelet for the possibility of trauma or illness. These patients have a loss of intrinsic cortisol production, and can suffer from cardiovascular collapse if their cortisol deficiency is not recognized and supplemented with stress-dose steroids.

Treatment for hypersecreting tumors depends on the hormone being secreted. It is very important to establish the hormonal profile of a tumor prior to treatment for this reason. Currently, prolactinomas are treated first with a dopamine agonist. The most common medications used to treat prolactinomas are cabergoline and bromocriptine. Cabergoline is a selective D-2 dopamine receptor agonist that is widely considered first-line therapy for prolactinomas. It is favored over bromocriptine because of its selectivity, twice weekly dosing, better tumor control, effective lowering of prolactin levels, greater return of gonadal function/menstrual cycles, and possibility of obviating the need for life-long therapy. Two randomized trials showed better prolactin normalization with cabergoline as compared to bromocriptine. Recently, association of cabergoline with cardiac valvular disease has prompted a reexamination of the initial treatment of prolactinoma, but currently medical treatment as first line is still favored. Cardiac complications are generally seen with much higher doses than required to treat a prolactinoma. Most prolactinomas respond well to medical treatment, but 10%-20% of patients do not. If medical therapy fails to reduce the tumor size or its mass effect, or sufficiently control prolactin levels, surgery via transsphenoidal approach may be offered. Surgical management occasionally may be indicated in patients with rapidly worsening visual loss, and can be considered in patients attempting pregnancy, although cabergoline is still the first-line intervention in these patients.

Other hypersecreting tumors are treated via transsphenoidal approach, as with nonsecreting adenomas. Postoperative laboratory testing confirms the efficacy of therapy. As with nonsecreting tumors, craniotomy and/or radiation are sometimes necessary, depending on the extent and location of tumor growth. Pituitary replacement may be necessary postoperatively as well.

In the case of pituitary apoplexy, treatment often consists of urgent transsphenoidal decompression to prevent further visual deficit, and prompt initiation of stress dose steroid replacement in order to avoid a pituitary crisis.

Summary

Pituitary adenomas are largely benign tumors that are either secreting or nonsecreting. Diagnosis is established with history, physical examination, MRI, and hormonal testing. Treatment consists of conservative therapy for nonsecreting micro/macroadenomas without symptoms of compression, medical therapy for prolactinomas, and transsphenoidal surgery for symptomatic macroadenomas and other hypersecreting tumors. Occasionally, craniotomy and/or radiation may be necessary for further tumor control. Depending on hormonal status, chronic pituitary replacement may be necessary postoperatively. Successful treatment of pituitary tumors can often be achieved with the combined effort of a medical endocrinologist and pituitary neurosurgeon.

Chandler WF: Treatment of disorders of the pituitary gland: pearls and pitfalls from 30 years of experience. Clin Neurosurg 2009;56:18-22.

Fernández-Balsells MM, Murad MH, Barwise A, Gallegos-Orozco JF, Paul A, Lane MA, Lampropulos JF, Natividad I, Perestelo-Pérez L, Ponce de León-Lovatón PG, Erwin PJ, Carey J, Montori VM: Natural history of nonfunctioning pituitary adenomas and incidentalomas: a systematic review and metaanalysis. J Clin Endocrinol Metab 2011 Apr;96(4):905-912.

Lania A, Beck-Peccoz P: Pituitary incidentalomas. Best Pract Res Clin Endocrinol Metab 2012 Aug;26(4):395-403.

Lucas JW, Zada G: Endoscopic surgery for pituitary tumors. Neurosurg Clin N Am 2012 Oct;23(4):555-569.

Oh MC, Kunwar S, Blevins L, Aghi MK: Medical versus surgical management of prolactinomas. Neurosurg Clin N Am 2012 Oct;23(4):669-678.

Sarwar K, Huda M, Van de Velde V, Hopkins L, Luck S, Preston R, McGowan B, Carroll P, Powrie J: The prevalence and natural history of pituitary haemorrhage in prolactinoma. J Clin Endocrinol Metab 2013 Jun;98(6):2362-2367.

Sheth SA, Bourne SK, Tritos NA, Swearingen B: Neurosurgical treatment of Cushing disease. Neurosurg Clin N Am 2012 Oct;23(4):639-651.

Steffensen C, Maegbaek ML, Laurberg P, Andersen M, Kistorp CM, Norrelund H, Sørensen HT, Jorgensen JO: Heart valve disease among patients with hyperprolactinemia: a nationwide population-based cohort study. J Clin Endocrinol Metab 2012 May;97(5):1629-1634.

Wind JJ, Lonser RR, Nieman LK, Devroom HL, Chang R, Oldfield EH: The lateralization accuracy of inferior petrosal sinus sampling in 501 patients with Cushing’s disease. J Clin Endocrinol Metab2013 Jun;98(6):2285-2293.

PEDIATRIC NEUROSURGERY

Debbie K Song, MD Cormac O. Maher, MD Karin M. Muraszko, MD

CONGENITAL MALFORMATIONS

Craniospinal Dysraphism

Craniospinal dysraphism results from improper formation and closure of the neural tube during development. These may be classified based on whether they are open or closed neural tube defects, the location of the lesion, or the embryological basis for the malformation. Open neural tube defects are those in which neural elements are exposed or covered by a dysplastic membrane, while closed neural tube defects are skin covered.

Myelomeningocele, the most common type of spinal dysraphism compatible with life, occurs with an incidence of 1 in every 1200-1400 live births. It is due to a local failure of neural tube closure during primary neurulation. Fusion of the lateral cutaneous ectoderm and the process of disjunction also fail to occur in myelomeningocele, resulting in a midline cutaneous defect over exposed neural tissue called the neural placode. Therefore, myelomeningoceles are considered open neural tube defects. Myelomeningoceles occur most commonly in the lumbar spine, and the anatomic level of the spinal cord lesion approximates the patient’s neurologic deficits. The diagnosis of a neural tube defect can be suspected prenatally with an elevated maternal serum alpha-fetoprotein and confirmed by in utero imaging such as a maternal-fetal MRI or ultrasound. Pregnant mothers who have inadequate folate intake or who have other children with neural tube defects are at increased risk for giving birth to a child with a myelomeningocele. A Chiari II malformation is found in most patients with myelomeningocele. Eighty percent of patients with myelomeningocele have associated hydrocephalus. Other CNS abnormalities that can be found with increased incidence among patients with myelomeningocele include lipomas, syringomyelia, and diastematomyelia. Patients with myelomeningocele commonly have orthopedic problems that include scoliosis, hip dislocation, knee, and foot deformities. Besides a neurogenic bladder, patients with myelomeningocele are at increased risk of genitourinary abnormalities, as well as intestinal, cardiac, esophageal, and renal abnormalities. Workup for the newborn child with myelomeningocele includes a cranial and spinal ultrasound, and orthopedic and urology consults. The neonate should be placed in the prone position, with pressure off of the myelomeningocele. The myelomeningocele should be covered in moist. Surgical closure of the myelomeningocele is performed soon after birth. The Management of Myelomeningocele Study is an ongoing trial examining the utility of in utero myelomeningocele repair.

Closed neural tube defects, also referred to as occult spinal dysraphisms, can arise from problems with disjunction, secondary neurulation, or postneurulation events. Types of closed neural tube defects include dermal sinus tracts, spinal lipomas, neurenteric cysts, sacral dysgenesis, and diastematomyelia. Spina bifida occulta, which is characterized by a defect in the posterior elements of the spine, is often a harbinger of an underlying closed neural tube defect. These malformations can tether the spinal cord in an abnormally low position and produce excessive tension on the neural elements. Neuronal dysfunction may ensue with symptoms of a clinical tethered cord syndrome such as back or leg pain, worsening lower extremity motor and sensory function, decline in bladder and bowel function, worsening lower extremity orthopedic deformities, and progressive scoliosis. Early surgical repair of occult spinal dysraphisms is often recommended at the time of diagnosis in order to prevent the onset or halt the progression of neurological symptoms.

Dermal sinus tracts are epithelial-lined tracts that originate in the midline skin, usually in the caudal lumbosacral region above the S2 level. The sinus tract extends from a pinhole opening in the skin, through bifid spinous processes, and extends into the dura to communicate with the spinal cord. The lining of the dermal sinus tract contains normal skin appendages which can shed and communicate with the intradural space, and recurrent episodes of meningitis and arachnoiditis may ensue. They appear as dimples above the gluteal crease and must be differentiated from pilonidal cysts, which are closer to the anus. Both entities may drain from the skin. The skin around the ostium of a dermal sinus tract may be discolored or have a hairy tuft. Dermal sinuses can be associated with lipomas, dermoid tumors, or epidermoid cysts at any point along the tract or within the spinal canal. Examination of the child with a suspected dermal sinus tract should include assessment of sphincter function, lower extremity reflexes, and motor and sensory function. Treatment should be performed in an expeditious fashion after diagnosis in order to reduce the risk of CNS infection and prevent the development of neurological deficit. Less commonly, dermal sinus tracts can occur in the cranial region. The most common cranial locations are in the occipital or nasal region. Children may present with a midline dimple at the tip of the nose or in the occipital region and a history of recurrent meningitis. Cranial dermal sinus tracts can be associated with intracranial dermoid cysts.

Spinal lipomas are the most common closed neural tube defects and include three separate entities: intradural lipomas, lipomyelomeningoceles, and lipomas derived from the caudal cell mass including fibrolipomas of the filum terminale. A lipomyelomeningocele consists of an intradural lipoma that is attached to the spinal cord and extends through defects in the dura, bony spine, and fascia to become continuous with the subcutaneous fat. Seventy percent of lipomyelomeningoceles are associated with subcutaneous fatty masses. Lipomyelomeningoceles present as skin-covered lumbosacral masses above the gluteal crease. The overlying skin may be discolored from a port-wine stain or hemangioma, have a hairy tuft, or contain an ostium of a dermal sinus tract. The caudal end of the spinal cord is usually tethered in a low-lying position in cases of lipomyelomeningocele, with the conus medullaris positioned below the normal L1-2 level. Filum terminale fibrolipomas and distal conus lipomas, in contrast to lipomyelomeningocles, are malformations of secondary neurulation. Intradural lipomas, lipomyelomeningoceles, and filum terminale lipomas can all tether the spinal cord. The neurological examination in such children may be normal, or patients may present with symptoms of a clinical tethered cord syndrome. Symptoms may become more prominent and neurological deficits can worsen during growth spurts. As a child gains weight, intraspinal lipomas will also undergo fat deposition which can compress or tether neural elements. Treatment of lipomyelomeningoceles includes cord untethering with resection or debulking of the intraspinal lipoma.

Diastematomyelia, also known as a split cord malformation, occurs when the spinal cord is split into 2 hemicords. The hemicords may be contained within separate dural sleeves separated by a bony septum, or both hemicords may be contained within a single dural sac and separated by a fibrous septum. The 2 hemicords reunite below the level of the lesion. Diastematomyelia is most commonly found in the lumbar spine and has a gender predilection for women. Children with diastematomyelia often have cutaneous stigmata such a nevus or a hairy tuft (hypertrichosis) at the level of the malformation. Bony anomalies, including spina bifida occulta, hemivertebrae, butterfly vertebrae, bony spurs at the level of the lesion, scoliosis, and orthopedic foot deformities are associated with split cord malformations. Clinically, diastematomyelia presents with symptoms of a tethered cord. Surgical treatment involves resection of any bony spurs and/or septum, untethering of the spinal cord, and reconstitution of a single dural sac.

Encephaloceles occur when there is a herniation of brain tissue and meninges through defects in the cranial vault. The tissue contained within an encephalocele consists of dysplastic and nonfunctional neural tissue with variable amounts of blood vessels, choroid plexus, dura, and ventricular tissue. The prognosis in children with encephaloceles is dependent on the amount of neural tissue contained within the encephalocele. Encephaloceles can be categorized as either posterior or anterior cranial fossa malformations depending on their location, and they are further classified according to the bone through which the herniation of tissue occurs. Posterior encephaloceles are associated with other midline congential anomalies, including myelomeningocele, Dandy–Walker malformation, Klippel–Feil anomaly, dorsal interhemispheric cysts, abnormalities of the corpus callosum, and neuronal migration disorders. For occipital and parietal encephaloceles, it is important to delineate the relationship of the lesion with adjacent venous sinuses. The goals of operative repair include removal of the encephalocele sac, preservation of any possible functional neural tissues, and closure of the dura in a water-tight fashion and of the wound with nondysplastic skin. Up to 50% of infants will develop hydrocephalus within 1 month of encephalocele repair; thus, surveillance with serial cranial ultrasounds is warranted in this group.

Arachnoid Cysts

Arachnoid cysts are developmental anomalies that form between separated layers of the arachnoid membrane. The walls of an arachnoid cyst may thicken with collagen deposition over time or hemorrhage. These cysts most commonly occur in the middle cranial fossa and retrocerebellar region. The brain may be shifted by the arachnoid cyst, but the overall brain volume is usually close to normal. Arachnoid cysts are asymptomatic, incidental findings in the vast majority of cases. Occasionally, they may present with symptoms specific to the location of the lesion. The natural history of an arachnoid cyst is usually benign. Most cysts do not change over time, although some may enlarge, shrink, or even disappear completely. Treatment is rarely indicated, and only if the arachnoid cyst is clearly symptomatic. There are various surgical treatment options available including endoscopic or open fenestration of the cyst as well as cystoperitoneal shunting.

Chiari Malformations

Chiari malformations consist of 4 types of congenital hindbrain abnormalities. Only Chiari Malformation Types I and II are commonly seen. Chiari I malformations are defined on imaging when the cerebellar tonsils extend at least 5 mm below the foramen magnum (Figure 36–18). The tonsils may assume a pointed, peg-shape instead of the normal rounded shape, causing crowding at the foramen magnum which limits CSF flow through the craniovertebral junction. Other CNS abnormalities associated with Chiari I malformation include syringomyelia (cavitation within the spinal cord), basilar invagination, platybasia, and Klippel–Feil anomaly. Children with Chiari I malformations are often asymptomatic but may also present with occipital headaches that are exacerbated by straining or coughing. Other symptoms include weakness, numbness, progressive scoliosis, long tract signs, central sleep apnea, or hydrocephalus. The work-up should include imaging of the brain and spine to assess for the presence of a syrinx and CSF flow studies to evaluate flow at the foramen magnum. For symptomatic children with a Chiari I malformation and syrinx, surgical treatment is recommended. Surgery is never indicated for asymptomatic children with no syrinx. The decision to proceed with surgical intervention in a child with a diagnosis of type I Chiari malformation on imaging, headaches, but no syrinx must be made carefully given the common occurrence of headaches as well as incidental Chiari I in the neurologically normal population. Surgical treatment entails some combination of a suboccipital craniectomy, C1 laminectomy, and duraplasty. Syringomyelia associated with Chiari I malformation usually resolves or improves significantly after posterior fossa decompression.

Figure 36–18. Sagittal T2-weighted MRI scan of a Chiari showing typical peg-like appearance of cerebellar tonsils and associated syringomyelia.

Chiari II malformations are found in patients with myelomeningocele. On imaging, these patients exhibit a low position of the cerebellar vermis and tonsils below the foramen magnum; elongation, kinking, and displacement of the medulla below the foramen magnum and around the cervical spinal cord; abnormal lamination of the cerebral cortex; and often an upward displacement of the rostral cerebellum through a low-lying tentorium. Additional associated MR imaging characteristics of Chiari II malformations include hydrocephalus, fusion of the inferior colliculi of the brainstem, a large thalamic massa intermedia, a high-riding third ventricle, an elongated fourth ventricle, and a disproportionately small posterior fossa with asymmetric and flattened cerebellar folia. Symptomatic patients may present with apnea and other respiratory abnormalities secondary to compression of the medullary respiratory control center. Long tract signs, headache, ataxia, and gait instability may also be evident. In addition to myelomeningocele, other CNS abnormalities associated with Chiari II malformations include basilar impression, corpus callosum abnormalities, and cortical malformations. For symptomatic patients, surgical treatment consists of a posterior fossa decompression. Chiari III and IV malformations are very rare. Chiari III malformations include features of Chiari II malformation as well as an occipital encephalocele. Chiari IV malformations are characterized by severe cerebellar hypoplasia without an encephalocele.

Craniosynostosis

Craniosynostosis refers to the premature fusion of one or more cranial sutures. When this occurs, bone growth is restricted in a direction perpendicular to the fused suture, and there is compensatory growth at other sites. The result is a misshapen head which may take one of several forms depending on which of the cranial sutures is involved. One or multiple sutures may be affected. The incidence of craniosynostosis is approximately 5 per 10,000 live births, and men are affected more than women.

Sagittal synostosis, the most common type of single suture synostosis, results in an elongated, boat-shaped skull referred to as scaphocephaly. In scaphocephaly, the biparietal diameter is reduced, while the anterior-posterior (AP) diameter is increased. Frontal bossing is common in this condition. Patients with sagittal synostosis have a palpable keel-like prominence over the fused sagittal suture.

Coronal synostosis may be either unilateral or bilateral. Unilateral coronal synostosis produces an asymmetric head shape known as plagiocephaly. The forehead on the affected side is flattened, while the forehead on the unaffected appears to bulge abnormally. Bilateral coronal synostosis results in brachycephaly, characterized by a broad, flattened forehead. The AP diameter is reduced, while the bitemporal and biparietal diameters are increased. When bilateral coronal synostosis occurs in combination with sagittal synostosis, turricephaly results. Turricephaly is characterized by a high, tower-like head shape with a vertical forehead.

Metopic synostosis is associated with trigonocephaly, a head shape that is characterized by a triangular-shaped forehead and hypotelorism. The bitemporal diameter is narrowed, and there is often a bony ridge in the midline of the forehead, over the fused metopic suture.

True unilateral lambdoid craniosynostosis is rare with an incidence of 1 in 300,000 live births, and it must be differentiated from positional posterior plagiocephaly, an increasingly common diagnosis. In unilateral lambdoid synostosis, there may be slight prominence of the forehead on the unaffected side. The ear on the affected side will be posteriorly and inferiorly displaced relative to the contralateral ear of the unaffected side. The head shape is trapezoidal when viewed from above.

While most cases of craniosynostosis are sporadic and involve a single suture, multiple-suture craniosynostosis occurs in certain genetic syndromes. Crouzon’s syndrome is an autosomal dominant disorder characterized by the premature fusion of the bilateral coronal, frontosphenoid, and frontoethmoid sutures. Clinically, the condition is characterized by brachycephaly, maxillary hypoplasia, shallow orbits, proptosis, and a beaked nose. Apert’s syndrome is an autosomal dominant condition which is characterized by pansynostosis. Clinically, patients have hypertelorism, midface hypoplasia, and shallow orbits in addition to their craniosynostosis. Symmetric syndactyly and short thumbs are also characteristic of Apert’s syndrome. Hydrocephalus is common in this condition.

Operative repair of craniosynostosis is often improving cosmesis. Operative approaches vary from endoscopic strip craniectomies of the involved suture to more extensive cranial vault reconstruction.

HYDROCEPHALUS

Disturbances in cerebrospinal fluid (CSF) circulation or absorption result in hydrocephalus. Hydrocephalus can be classified into 2 types: obstructive or communicating. In obstructive hydrocephalus, CSF circulation is blocked within the ventricular system and there is enlargement in the ventricles proximal to the obstruction. In communicating hydrocephalus, CSF absorption is blocked at the level of the arachnoid granulations. Rarely, hydrocephalus may be due to the overproduction of CSF, as is the case in certain choroid plexus tumors.

The incidence of congenital hydrocephalus ranges from 0.9 to 1.8 per 1000 births. Neonatal hemorrhages of the germinal matrix and choroid plexus, as well as infections can cause adhesions to form in the cerebral aqueduct or at the foramen of Magendie and Luschka, interfering with CSF absorption.

Hydrocephalus can cause elevations in intracranial pressure which may manifest in different ways depending on the age of the child. In neonates and infants whose anterior fontanelle is still open, untreated hydrocephalus will present with a tense or bulging fontanelle, apneic and bradycardic episodes, engorgement of the scalp veins, upward gaze palsy, gaps between the cranial sutures, rapid increases in head circumference, irritability, poor head control, and poor oral intake. In children with a closed cranial vault whose fontanelle has closed, untreated hydrocephalus will present with symptoms of intracranial hypertension including lethargy or excessive sleepiness, papilledema, headache, nausea, vomiting, gait disturbance, increased fussiness, or upgaze or lateral gaze palsy.

Several surgical options can be considered in the treatment of hydrocephalus. The most common CSF diversionary procedure is ventriculoperitoneal shunting, creating a shunt between the cerebral ventricles and the peritoneal cavity. Other types of shunts may drain into other locations including the right atrium (ventriculoatrial shunt) or pleural cavity (ventriculopleural shunt). In children with certain types of obstructive hydrocephalus, an endoscopic third ventriculostomy may be considered, which involves fenestration of the floor of the third ventricle, thereby creating an alternative CSF pathway.

Shunt failure or infection may manifest with signs and symptoms of acute intracranial hypertension. Ventricular enlargement may or may not be present in shunt failure. Prompt treatment of acute hydrocephalus and/or shunt failure is indicated to prevent irreversible neurologic injury including herniation, blindness, or death.

PEDIATRIC CENTRAL NERVOUS SYSTEM TUMORS

Brain tumors are the most common solid tumors of childhood. The locations and types of tumors in the pediatric population differ from those in adults. Approximately two-thirds of brain tumors in children between 2 and 12 years of age occur in the infratentorial space. Brain tumors present in varying manners among different age groups. In neonates and infants, brain tumors may manifest with nonspecific findings, and mass effect from a tumor may not be clinically evident initially due to a compliant skull and an open fontanelle. In young children, a primary brain tumor may present with symptoms related to intracranial hypertension such as headache, nausea, and vomiting. Papilledema may be evident on funduscopic examination. Older children more often present with focal neurological signs and symptoms. The most common pediatric posterior fossa brain tumors are medulloblastoma, juvenile pilocytic astrocytoma, and ependymoma. When a posterior fossa brain tumor is diagnosed, preoperative imaging of the entire neuraxis should be performed whenever possible to evaluate for drop metastases in the spinal canal.

Medulloblastoma

Medulloblastomas comprise approximately 20% of all pediatric brain tumors and 30% of all posterior fossa tumors in children. They appear as hyperdense lesions in the region of the fourth ventricle on CT imaging and enhance following contrast administration (Figure 36–19). Complete or near-complete surgical resection is the goal, as the extent of residual tumor is related to prognosis. Patients with medulloblastoma are stratified into either a standard or high-risk group. Those children who are younger than three years, have greater than 1.5 cm2 residual tumor on postoperative imaging, or have dissemination of tumor away from the primary site as deemed by either imaging studies or positive CSF cytology have a worse prognosis and are classified as high risk. Postoperative craniospinal radiation with a boost to the posterior fossa is indicated in children older than 3 years. Those patients who are classified as high risk are typically treated with chemotherapy as well. Because of the increased morbidity of radiation in children less than 3 years of age, chemotherapy is used to delay the radiation dose in such young patients. Recurrences, if they occur, typically occur within 3 years. The 5-year survival in standard risk patients is 70%, while that in high-risk patients is approximately 40%. Up to 25% of patients with a posterior fossa medulloblastoma may require a CSF diversionary procedure due to persistent postoperative hydrocephalus.

Figure 36–19. Head CT demonstrating large mass (medulloblastoma) within the fourth ventricle causing ventricular dilation.

Cerebellar Astrocytoma

Cerebellar astrocytomas account for approximately 20% of all pediatric brain tumors. The peak age of presentation is 10 years. The characteristic appearance on imaging studies is an enhancing mural nodule with a surrounding cyst. The goal of treatment is complete surgical resection, as patients with a gross total resection have a 90% long-term survival rate without any additional adjuvant therapies.

Ependymoma

Ependymomas can arise anywhere along the neuraxis in relation to an ependymal surface. In the pediatric population, 90% of ependymomas are intracranial, and of these, two-thirds are located in the posterior fossa. Ependymomas commonly arise from the floor of the fourth ventricle in close proximity to the brainstem. When located in the fourth ventricle, ependymomas may extend out the foramina of Luschka and Magendie into the surrounding CSF subarachnoid cisterns. Dissemination of ependymoma tumor cells in the CSF can occur in up to 10% of cases, underscoring the importance of full neuraxis imaging to properly stage the disease. Treatment usually consists of tumor resection and postoperative focal radiation.

Brainstem Glioma

Brainstem gliomas are a heterogeneous group of tumors of varying histology, biological behavior, and prognosis. Brainstem gliomas are typically be subdivided into four groups based on imaging characteristics: diffuse brainstem gliomas, focal brainstem gliomas, dorsally exophytic brainstem gliomas, and cervicomedullary gliomas. Diffuse brainstem gliomas represent up to 80% of all brainstem gliomas and carry the worst prognosis. They most commonly occur in the pons and can extend into the medulla or midbrain. The majority of children who are affected are between 6 and 10 years of age, and they present with a relatively short clinical history of unilateral or bilateral cranial neuropathies, progressive ataxia, gait abnormality, and long tract signs. On MR imaging studies, diffuse brainstem gliomas appear as non-enhancing, hypointense masses that expand the pons (Figure 36–20). They appear hyperintense on T2-weighted sequences, and with disease progression the tumor may completely encase the basilar artery. Histologically, they are malignant (WHO grade III or IV) astrocytomas. Diagnosis of a diffuse brainstem tumor can be made on the basis of imaging alone, and biopsy is usually not recommended. Radiation therapy and steroids may improve symptoms but have not been shown to prolong survival. These tumors are universally fatal, with a median survival of 8-10 months.

Figure 36–20. Magnetic resonance imaging appearance of a diffuse pontine glioma.

Other types of brainstem gliomas are associated with a better prognosis. Dorsally exophytic brainstem tumors grow from the subependymal surface into the fourth ventricle, away from the brainstem. They are characterized by slow growth with gradual onset of symptoms. Eventually they may obstruct CSF outflow from the fourth ventricle and result in hydrocephalus. Surgical excision should be performed when the predicted morbidity is not prohibitive. Histologically, focal intrinsic and dorsally exophytic brainstem gliomas are usually lower grade (WHO grade I or II) lesions. Cervicomedullary brainstem gliomas have similar behavior and histology as intramedullary spinal cord gliomas. These tumors may cause symptoms including weakness and lower cranial neuropathies. They should be resected.

Tumors of Infancy

Brain tumors are present in 1.1 newborns per 100,000 births. Infantile tumors include medulloblastoma, central neuroblastoma, supratentorial primitive neuroectodermal tumor (PNET), pineoblastoma, and atypical teratoid/rhabdoid tumor (AT/RT). These are all high-grade tumors with a propensity to spread throughout the CSF. Medulloblastoma, central neuroblastoma, supratentorial PNET, and pineoblastomas are all considered PNETs with similar histologies. The combination of bilateral retinoblastomas and a midline pineoblastoma is referred to as “trilateral neuroblastoma” and carries a dismal prognosis. Atypical teratoid/rhabdoid tumors are usually found in the posterior fossa and are associated with deletions on chromosome 22 in over 90% of cases. AT/RTs have a poor prognosis, and most children die within 1 year of diagnosis.

Pineal Region Tumors

Tumors arising in the region of the pineal gland comprise 3%-8% of pediatric brain tumors. Histologically, pineal region tumors are most often germ cell tumors such as germinomas, teratomas (mature and immature), embryonal cell carcinomas, choriocarcinomas, and endodermal sinus tumors. Germinomas are the most common pineal region tumors and demonstrate a gender predilection for men. Pineal parenchymal tumors such as pineocytoma or pineoblastomas occur less frequently. Pineal tumors can compress the cerebral aqueduct and cause hydrocephalus. Patients may present with a Parinaud’s syndrome consisting of impaired upgaze, convergence-retraction nystagmus, lid retraction, convergence paralysis, pupillary dilatation, and light-near dissociation. Complete neuraxis imaging should be performed as drop metastases can occur via CSF pathways. Serum and CSF should be tested for tumor markers that may be secreted by germ cell tumors including placental alkaline phosphatase, alpha-fetoprotein, and beta-human chorionic gonadootropin (beta-HCG). Radiation therapy with or without chemotherapy is the mainstay of treatment for germ cells tumors.

DNET

Dysembryoplastic neuroepithelial tumor (DNET) is a low-grade cortical based tumor with a median age of presentation of 7 years. They appear as superficial cystic tumors in the temporal or frontal lobes. On MR imaging, they are hypointense on T1-weighted sequences, hyperintense on T2-weighted sequences, and do not enhance. The affected cortical gyrus often has a bubbly appearance. There is no edema or mass effect associated with DNETs. Patients with these tumors typically present with a long history of intractable complex partial seizures and have a normal neurological exam. Gross total resection of the tumor is curative and usually eliminates seizures.

Sellar Tumors

Tumors that arise in the region of the sella turcica and pituitary gland in children include pituitary adenomas and craniopharyngiomas. Pituitary adenomas are relatively rare among children. Craniopharyngiomas, however, represent 6%-9% of all pediatric brain tumors and are the most common nonglial intracranial masses in children. On imaging, craniopharyngiomas appear as cystic tumors which originate from a suprasellar location. Calcification within these tumors is common. Craniopharyngiomas are hyperintense on T1- and T2-weighted sequences due to fat, cholesterol, and proteinaceous contents within the cystic tumor. Craniopharyngiomas may cause symptoms attributable to intracranial hypertension and hydrocephalus or they may present with visual disturbances. Also, craniopharyngiomas may cause endocrine disturbances such as growth failure, diabetes insipidus, hypothyroidism, or menstrual dysfunction. A thorough endocrine workup is warranted in patients with suspected craniopharyngioma. The surgical approach depends on the location and extent of the craniopharyngioma. Potential surgical complications include diabetes insipidus or hypothalamic insufficiency.

Hypothalamic Optic Gliomas

Gliomas of the optic pathway are more common in the pediatric population. Included in this category are optic nerve gliomas and chiasmatic/hypothalamic astrocytomas. Histologically, these tumors are most often pilocytic astrocytomas. Optic nerve gliomas are associated with neurofibromatosis 1 (NF-1). Gliomas of the optic chiasm and hypothalamus are cystic, globular tumors that enhance and rarely calcify. They may cause hydrocephalus due to compression at the foramen of Monro. Infants with hypothalamic/chiasmal gliomas present with visual loss, macrocephaly, and a diencephalic syndrome consisting of failure to thrive, cachexia, motor hyperactivity, and hyperalertness. Children 2-5 years of age may present with visual loss and endocrine dysfunction including short stature or precocious puberty. Symptoms in older children include visual loss and hypopituitarism. The goals of surgical treatment for chiasmatic/hypothalamic tumors are to obtain a tissue diagnosis and reestablish patent CSF pathways. Radiation has a clear benefit in extending progression-free survival. The 10-year relapse-free rate is approximately 55% with radiation versus 14% with biopsy alone. Chemotherapy is reserved for younger patients less than 5 years of age in order to delay radiation. The 10-year survival following surgical biopsy and radiation for chiasmatic/hypothalamic gliomas ranges from 48% to 55%; however, these tumors and their treatments are associated with significant morbidity, including visual impairment, endocrine dysfunction, obesity, and neurocognitive decline.

Choroid Plexus Tumors

Choroid plexus tumors include choroid plexus papillomas and choroid plexus carcinomas. They represent 2%-4% of all pediatric brain tumors. Choroid plexus papillomas occur in the atrium of the lateral ventricle in children and are attached to normal choroid plexus. These tumors avidly enhance and may calcify. If a gross total resection of the tumor is achieved, no adjuvant therapy is required. Choroid plexus papillomas are WHO grade I or II tumors and have a good prognosis. Choroid plexus carcinomas are malignant tumors for which the average age of diagnosis is 2 years. Like choroid plexus papillomas, these tumors are usually located in the lateral ventricles. Forty-five percent of choroid plexus carcinomas demonstrate dissemination at diagnosis. These tumors contain necrosis and can hemorrhage. Treatment consists of surgical resection, radiation therapy, and possibly chemotherapy. For children younger than 3 years of age, multiagent chemotherapy is used to delay the onset of radiation therapy. The prognosis is poor.

Spinal Cord Tumors

Spinal cord tumors in children account for 15% of all pediatric CNS tumors. Spinal cord tumors typically present with progressive back and leg pain, neurologic deficit, gait instability, torticollis, or bowel and bladder dysfunction. The most common intradural-intramedullary spinal cord tumors are low-grade glial tumors including astrocytomas. Treatment consists of early surgery, as postoperative morbidity is worse if preoperative neurologic deficits exist. Near-total resections of low-grade spinal cord gliomas in children can confer long-term progression-free survival. High-grade gliomas are treated with surgical debulking followed by adjuvant therapies. Intradural-extramedullary tumors in children include dermoid cysts, teratomas, and neurofibromas. Extradural primary spinal tumors that occur in childhood can present with myelopathy and spinal cord compression; such tumors include aneurysmal bone cysts, osteoid osteomas, osteoblastomas, eosinophilic granulomas, and rarely, metastatic disease.

Phakomatoses

Phakomatoses are neurocutaneous syndromes that manifest with skin lesions and central nervous system tumors. Most of the phakomatoses are inherited conditions. NF-1 is an autosomal dominant inherited condition caused by a mutation on chromosome 17. Optic nerve gliomas are associated with NF-1 and, in affected children, usually occur before 6 years of age. Neurofibromatosis 2 (NF-2) is an autosomal dominant syndrome that results from a mutation on chromosome 22. Typical CNS tumors associated with NF-2 include bilateral acoustic neuromas, meningiomas, schwannomas, and intramedullary spinal cord ependymomas.

Tuberous sclerosis (TS) is an autosomal dominant syndrome that arises from mutations in chromosomes 9, 11, or 16. Children with TS can get periventricular hamartomas known as subependymal nodules near the foramen of Monro and adjacent to the caudate nucleus. In 15% of patients with TS, subependymal nodules can transform into a subependymal giant cell astrocytoma, a WHO grade I lesion. These are benign, enhancing tumors that arise at the foramen of Monro and cause obstructive hydrocephalus. These tumors typically occur prior to the end of the second decade of life. They can enlarge with time, and gross total resection of the subependymal giant cell astrocytoma is considered curative. The CNS lesions in TS frequently cause seizures.

Von Hippel–Lindau (VHL) disease is an autosomal dominant disease due to a mutation on chromosome 3. Patients with VHL can get hemangioblastomas, most commonly in the posterior fossa and spinal cord. Although hemangioblastomas are considered benign tumors, they may recur in multiple locations in patients with VHL.

CEREBROVASCULAR DISEASE IN CHILDREN

Aneurysms & Vascular Malformations

Children may present with a variety of intracranial vascular malformations such as arteriovenous malformations (AVMs), venous angiomas, capillary telangiectasias, and cavernous malformations. AVMs may present with seizures or focal deficits from a hemorrhage. Without treatment, patients are at risk for recurrent hemorrhages. AVMs are usually treated with surgery in the pediatric age group but may occasionally be treated with stereotactic radiation or embolization techniques. In patients with an autosomal dominant inherited cavernous malformation syndrome, the cavernous malformations may be multiple and hemorrhage at an earlier age. Cavernous malformations are treated with surgical resection. Intracranial saccular aneurysm rupture is rare in the pediatric population. Depending on the size and configuration of the aneurysm, these lesions may be treated by surgical clipping, endovascular coiling, or managed conservatively.

Vein of Galen Malformations

Vein of Galen malformations are congenital vascular malformations characterized by extensive arterial feeders draining into an enlarged vein of Galen. Although these malformations are also known as Vein of Galen aneurysms, they represent arteriovenous fistulae. Newborns can present with high-output cardiac failure due to the arteriovenous shunting. Hydrocephalus is common from compression of the cerebral aqueduct by the malformation. Seizures are also associated with these lesions. The extensive arteriovenous shunting can produce a steal effect and result in cerebral ischemia and infarction. The prognosis may be poor for patients diagnosed in early infancy with heart failure. The prognosis is better for those diagnosed later in life. Treatment usually involves endovascular embolization of feeding arteries.

Moyamoya Disease

Moyamoya disease is an idiopathic vasculopathy that leads to progressive occlusion of one or both internal carotid arteries with secondary formation of a collateral capillary network at the base of the brain. The disease can also involve the proximal middle and anterior cerebral arteries. On angiography, the collateral vessels have a characteristic “puff-of-smoke” appearance. Children with Moyamoya disease present with ischemic events that may be provoked by straining or hyperventilation. Refractory headaches, seizures, and alternating hemiplegia are also associated with Moyamoya disease. Moyamoya is treated with surgical revascularization to improve blood flow via direct or indirect bypass procedures.

SPASTICITY

Spasticity in children is most often due to cerebral palsy, and several surgical options are available for treatment. In determining whether a child with hypertonia will benefit from surgical intervention, it is important to assess if dystonia is present as well and its contribution to the hypertonia, the ambulatory potential of the child, and to what extent the underlying spasticity is useful for the child in terms of providing strength and allowing them to support their own weight. In addition to medication, orthopedic procedures, and periodic injections that are used to treat spasticity, the neurosurgical treatments for spasticity include placement of an intrathecal baclofen pump and selective dorsal rhizotomy (SDR). An intrathecal baclofen pump entails inserting a catheter into the intrathecal space and connecting it to a subcutaneous pump so that the antispasticity drug baclofen may be continuously delivered. Depending on at what spinal level the catheter tip is placed, upper and lower extremity spasticity may be treated. Intrathecal baclofen pumps are useful if the upper extremities are affected by severe hypertonia, if the tone conferred by spasticity is required for standing or walking, or if the lower extremity spasticity in nonambulatory patients is disabling and hinders care of the patients. Those patients who are ambulatory and whose spasticity primarily affects their lower extremities may be candidates for an SDR. It is thought that inputs entering the spinal cord through the dorsal roots have a net excitatory effect on the anterior roots, thus contributing to spasticity. The premise of an SDR is to intraoperatively stimulate lumbosacral dorsal nerve rootlets and record responses from the anterior nerve roots and muscles. This allows for the identification of those nerve rootlets that are relatively more involved into maintaining hypertonia, and such nerve rootlets are sectioned. SDR has been shown to improve ambulation, but the procedure does not confer previously nonambulatory patients the ability to walk.

PEDIATRIC TRAUMA AND BIRTH INJURIES

General Principles

Head injury and its management are discussed elsewhere, and treatment principles used in the management of adult trauma also apply to the pediatric patient. Certain aspects of trauma that are unique to the pediatric population are highlighted. Head injuries are 30 times more common in children than spinal cord injuries, and they represent the most common cause of mortality and morbidity in children. In infants and young children, the brain and head are disproportionately large compared to the trunk and torso, and the neck and paraspinal musculature are incompletely developed. In children younger than 4 years, the skull is soft, unilaminar, and without diploë; as a result, it provides less protection to the brain for absorbing a traumatic impact and is more prone to fracture. Skull fractures in children can be linear, depressed, or ping-pong ball fractures. Ping-pong ball fractures occur in newborns and appear as a focal area of caved in skull that resembles a crushed ping-pong ball. No surgical intervention is required for ping-pong ball fractures in the temporoparietal region, as the growing skull will correct the deformity. Surgical elevation of a frontal ping-pong ball fracture may be considered for cosmetic purposes.

Nonaccidental Trauma

Nonaccidental head trauma is the leading cause of death and morbidity in children less than 2 years of age. In shaken baby syndrome, there may be few signs of external trauma with significant neurological injury. Alternatively, the child may present with lethargy, irritability, poor feeding, apneic episodes, or seizures. Multiple skull fractures that are associated with underlying brain injury, bilateral chronic subdural hematomas or subdural hematomas of varying ages, subarachnoid hemorrhage, and retinal hemorrhages should raise the index of suspicion for child abuse. Subdural hemorrhages commonly occur along the bilateral convexities or in the posterior interhemispheric fissure. MR imaging is best at evaluating subdural hematomas of varying ages, as well as the extent of DIA that occurs with the acceleration-deceleration and rotational forces in shaken baby syndrome. A workup should include imaging of the brain and possibly spine, a skeletal survey to assess for long-bone or rib fractures, a funduscopic examination to check for retinal hemorrhages, and a thorough external exam to assess for bruising. Death from nonaccidental trauma is most often due to refractory intracranial hypertension.

Spine Trauma

Spinal cord injury is relatively rare in the pediatric population and accounts for 5% of all spinal cord injury. The pediatric spine continues to develop throughout the first 2 decades of life. Ligamentous injury is more common than bony injury, owing to ligamentous laxity, immature supporting musculature, and the developing bony joints of the spine. The cervical spine is most commonly injured in children, and in children less than 9 years of age, two-thirds of cervical spine injuries occur between the C1 and C3 levels.

Al-Holou WN, Yew AY, Boomsaad ZE, Garton HJ, Muraszko KM, Maher CO: Prevalence and natural history of arachnoid cysts in children. J Neurosurg Pediatr 2010;5(6):578-585.

Cunningham ML, Heike CL: Evaluation of the infant with an abnormal skull shape. Curr Opin Pediatr 2007;19:645-651.

Lew SM, Kothbauer KF: Tethered cord syndrome: an updated review. Pediatr Neurosurg 2007;43:236-248.

Nield LS, Brunner MD, Kamat D: The infant with a misshapen head. Clin Pediatr 2007;46:292-298.

Shu HG, Sall WF, Maity A, et al: Childhood intracranial ependymomas. Twenty-year experience from a single institution. Cancer 2007;110:432-441.

Strahle J, Muraszko KM, Kapurch J, Bapuraj JR, Garton HJ, Maher CO: Chiari malformation Type I and syrinx in children undergoing magnetic resonance imaging. J Neurosurg Pediatr 2011;8(2):205-213.

Strahle J, Muraszko KM, Kapurch J, Bapuraj JR, Garton HJ, Maher CO: Natural history of Chiari malformation type I following decision for conservative treatment. J Neurosurg Pediatr 2011:8(2):214-221.

INTRACRANIAL ANEURYSMS AND ARTERIOVENOUS MALFORMATIONS

Aditya S. Pandey, MD Tristram Horton, MD John A Cowan, Jr, MD Neeraj Chaudhary, MD Joseph G. Gemmete, MD B. Gregory Thompson, MD

INTRACRANIAL ANEURYSMS

General Considerations

Intracranial aneurysms (IAs) represent an abnormal dilation or expansion of an artery within the cranial vault. Autopsy studies suggest that IAs are present within 1%-5% of the population. Most epidemiologic studies suggest that CAAs are more common in women (3:2) and results in the clinical presentation of approximately 30,000-35,000 people annually in the United States. Aneurysms are categorized as saccular, fusiform, or mycotic, and can be ruptured, expanding, or unruptured. Aneurysms can be located anywhere within the cerebral artery tree but are most commonly located in the Circle of Willis. The specific type, status, and location of an IA can drastically affect a patient’s clinical presentation, treatment options, and outcome. Intracranial aneurysms can be detected with a variety of imaging modalities, including cerebral angiography (Figure 36–21), computed tomography angiography (CTA), or magnetic resonance angiography (MRA). Patients with intracranial aneurysms should be referred to a neurosurgeon who specializes in the treatment of neurovascular diseases.

Figure 36–21. An anterior view of left carotid artery cerebral angiogram demonstrating a large cerebral artery aneurysm (arrow) at the middle cerebral artery bifurcation.

Clinical Findings

A ruptured intracranial aneurysm typically presents as a sudden, severe headache associated with nuchal rigidity and lethargy. Patients often describe the headache as the “worst headache of my life,” however, sentinel (or smaller) bleeds may not present with such severity. Other features at presentation can include vomiting, seizure, focal neurologic deficit (eg hemiparesis, oculomotor palsy, etc), or coma. The Hunt–Hess grading scale is commonly used to convey the severity of symptoms and is often used to risk-stratify patients (Table 36–6). The Fisher and World Federation of Neurological Surgeons are other grading scales commonly employed for patients presenting with spontaneous subarachnoid hemorrhage (SAH), based radiographically and clinically, respectively. Approximately 10%-20% of patients die prior to reaching a hospital and the overall mortality rate approaches 30%-50%. The peak age of rupture is between 45 and 55 years. The rebleed rate of a ruptured intracranial aneurysm is approximately 25% at two weeks and 50% by 6 months. The mortality of a rebleed approaches 80%.

Table 36-6. Hunt–Hess scale.1

Since the cerebral arteries course within the subarachnoid space, aneurysm rupture results in SAHs. Subarachnoid hemorrhages have a classic appearance on computed tomography (CT) scans, where the blood fills the normal cerebrospinal fluid (CSF) spaces surround the cortex, brainstem, and cerebellum (Figure 36–22). The differential diagnosis of SAH includes aneurysm rupture, trauma, coagulopathy, pretruncal (or perimesencephalic) venous bleed, cranial/spinal AVM, or dural venous sinus thrombosis. Intracranial aneurysm rupture may cause intracerebral hemorrhage, subdural hemorrhage, and intraventricular hemorrhage.

Figure 36–22. Classic appearance of a large subarachnoid hemorrhage. Notice the hemorrhage pattern fills the cerebrospinal fluid spaces at the base of the brain and around the brainstem.

Expanding IAs may or may present with symptoms of SAH. Patients with an expanding IA may exhibit neurologic symptoms consistent with focal compression. Though often not as dramatic a presentation as with rupture, expanding CAAs should be treated as emergencies. The specific deficit at presentation is dependent on the location of the aneurysm. Aneurysms along the posterior communicating (PComm) artery, classically present with a third nerve palsy (pupil dilation, eye abducted, and downward in position). Aneurysms along the posterior cerebral artery, although rarer, can have such a presentation. Aneurysms on the anterior communicating (AComm) artery can present with signs of optic apparatus compression. Cavernous internal carotid artery aneurysms can produce oculomotor palsies with retro-orbital pain. Ophthalmic artery aneurysms can lead to unilateral vision loss. Giant aneurysms (> 2.5 cm) can lead to more pronounced symptoms, including hemiparesis, obstructed hydrocephalus, hypothalamic dysfunction, seizures, and brainstem compression due to mass effect.

With the increasing availability and resolution of neuroimaging modalities, the detection of unruptured (and often asymptomatic) aneurysms is increasing. Although controversial, the estimated rupture rate of an unruptured intracranial aneurysms is 0.1%-2% per year. The rupture risk is dependent upon a number of factors, including aneurysm size, shape, and location. Other factors, including the hypertension and smoking history, as well as genetic predisposition (female sex, family history of SAH, certain arteriopathies), also play an important role in estimating the risk of aneurysm growth and rupture. Given the potentially devastating consequences of IA rupture, however, treatment of an unruptured IA should be considered in most cases.

Treatment

  1. Initial Management

A rapid history and physical examination of the patient including determination of the patient’s airway, breathing, and circulation (“ABCs”) should be performed in patients with a suspected ruptured intracranial aneurysm. Patients with unstable or unprotected airways or poor respiratory effort should be intubated and placed on mechanical ventilation. Maintenance of Pco2 values between 28 and 32 mm Hg can help acutely lower intracranial pressure and this range promotes adequate cerebral perfusion. Blood pressure should be controlled with a ceiling level of 160/90 mm Hg, maintained as needed with intravenous medications. Prophylactic anticonvulsant and gastrointestinal (H2 blocker or proton pump inhibitor) should be given as well. An arterial line and central venous line should be placed to further assist in management. Basic laboratory values needed include blood gas, complete blood count, coagulation function, sodium level, blood urea nitrogen, and creatinine. An electrocardiogram and chest x-ray should also be performed. Occasionally, patients presenting with SAH can exhibit signs and symptoms of pulmonary edema and/or heart failure requiring further intervention. Determination of the type and status of the CAA, via cerebral imaging, is paramount in dictating the remainder of the treatment algorithm. Demonstration of SAH on the initial head CT in patients with suspected CAA should be immediately followed by an assessment of the cerebral vascular tree (conventional angiography, CTA, etc). If the head CT does not demonstrate SAH and suspicion is high, a lumbar puncture can be performed. Typically, in the setting of CT-negative SAH, a lumbar puncture will reveal xanthochromia or high red blood cell counts that do not decrease (or “clear”) across serial CSF samples. Cerebrospinal fluid diversion, typically through ventriculostomy, can be performed if the patient presents with hydrocephalus or has concerns for elevated intracranial pressure.

  1. Surgical Management

Currently, two (three) main modalities exist for treating CAAs: surgical clip occlusion and endovascular coiling, and more recently, endovascular flow diversion. Surgical clip occlusion requires an open craniotomy and microsurgical dissection to the base of a CAA in order to apply a surgical clip around the neck of the aneurysm. In some cases, the aneurysm may require trapping and bypass (either surgical or endovascular) for treatment. Endovascular coiling utilizes a similar approach as conventional angiography and is thus less invasive. The endovascular surgeon navigates a microcatheter to the aneurismal defect and inserts detachable coils into the aneurysm dome. These coils promote thrombus formation and thus exclude the aneurysm from the native circulation.

The decision between clipping or coiling a CAA is complex and beyond the scope of this text. Factors considered in this decision include a patient’s age, overall medical condition, and preference as well as the aneurysm’s size, location, and morphology and the ruptured versus unruptured status of the aneurysm. To date, one trial has attempted to compare the two approaches for a very select group of ruptured CAAs. The study concluded that endovascular coiling resulted in slightly less morbidity and mortality at one year follow-up. The three randomized prospective trials comparing clipping and coiling have all found coiling to have lower risk of unfavorable outcome compared to clipping in the treatment of ruptured intracranial aneurysms. One study found clipping to be superior to coiling, with a significantly higher increased risk of death or readmission secondary to rebleeding in patients treated endovascularly. However, this study was limited by its retrospective cohort design. Case series with long-term follow-up have demonstrated some increase in rebleeding and/or the need for further treatment with endovascular coiling as compare to clip occlusion. More recent studies have shown a decrease in the recurrence and rebleeding rate after endovascular treatment of intracranial aneurysms. One study found that rebleeding rate was a function of degree of aneurysm occlusion on initial treatment, rather than the treatment modality itself.

  1. Medical Management

The medical management of patients who undergo either surgical clip occlusion or endovascular coiling for a ruptured CAA is particularly challenging. Patients require recovery in an intensive care setting that has particular expertise in neurological conditions. Once an aneurysm has been secured the blood pressure parameters are loosened and “permissive” hypertension is allowed (typically do not treat unless SBP > 200 mm Hg). Early tracheostomy and enteral feeding tubes are placed in patients who have neurologic deficits affecting respiratory or swallowing function. Patients are typically placed on Nimodipine, which has been demonstrated to slightly decrease postoperative vasospasm. Magnesium sulfate infusions are used in some centers for prevention of vasospasm although the data for this, while promising, is emerging.

Vasospasm is an idiopathic response of cerebral blood vessels to subarachnoid blood whereby the vessel constricts thus limiting distal blood flow. The peak time for vasospasm occurs between 4 and 14 days postbleed. Approximately 20%-40% of patients with have symptomatic vasospasm with 30% of those suffering a permanent neurologic deficit. Vasospasm can occur anywhere along a vessel and in vessels distant from the treated aneurysm. Patients can exhibit significant neurologic sequelae from vasospasm such as hemiparesis, aphasia, visual disturbance, etc depending on the particular vessel affected. Subtle findings including elevation in temperature and mental status changes can be harbingers of vasospasm. Vasospasm is typically treated using a regimen referred to as “triple H” therapy. This therapy involves hypertension (using vasopressors if needed to achieve SBP > 180), hemodilution (achieve hematocrit ~30%), and hypervolemia (using albumin or hypertonic solutions to achieve CVP 8-14 mm Hg). Cerebral angioplasty is an effective means for treating proximal constriction and is a first line treatment for symptomatic patients. Distal or diffuse spasm can respond to injection of calcium-channel blockers (eg, verapamil and nicardipine) papaverine, or milrinone (a phosphodiesterase-3 inhibitor) through a super-selective microcatheter.

Outcomes/Prognosis

For elective surgical clip occlusion or endovascular coiling, mortality rates (1%-2%) and morbidity rates (5%-10%) are relatively low. Risk factors for poor outcome would include aneurysm location, size, and presence of intraoperative rupture as well as other comorbid conditions (eg, coronary artery disease, diabetes, age, etc). In patients presenting with SAH, patient age, comorbid conditions, and Hunt–Hess grade are the strongest predictors of outcome. Overall, for patients stable enough for surgical intervention, mortality rates range from 10% to 20% with morbidity rates of 20%-40%.

Amenta PS, Dalyai RT, Kung D, et al: Stent-assisted coiling of wide-necked aneurysms in the setting of acute subarachnoid hemorrhage: experience in 65 patients. Neurosurgery 2012;70(6):1415-1429.

International Subarachnoid Hemorrhage Aneurysm Trial (ISAT) of Neurosurgical Clipping versus Endovascular Coiling in 2143 Patients with Ruptured Intracranial Aneurysms: a randomised trial.Lancet 2002;360:1267-1274.

Johnston SC: Rates of delayed rebleeding from intracranial aneurysms are low after surgical and endovascular treatment. Stroke 2006;37:1437-1442.

Johnston SC, Dowd CF, Higashida RT, et al: Predictors of rehemorrhage after treatment of ruptured intracranial aneurysms. The Cerebral Aneurysm Rerupture after Treatment (CARAT) Study. Stroke2008;39:120-125.

Molyneux AJ, Kerr RS, Yu L, et al: International Subarachnoid Aneurysm Trial (ISAT) of Neurosurgical Clipping versus Endovascular Coiling in 2143 Patients with Ruptured Intracranial Aneurysms: a randomised comparison of effects on survival, dependency, seizures, rebleeding, subgroups, and aneurysm occlusion. Lancet 2005;366:809-817.

McDougall CG, Spetzler RF, Zabramski JM, et al: The barrow ruptured aneurysm trial. J Neurosurg 2012;116:135-144.

O’Kelly, CJ, Kulkarni AV, Austin PC, et al: The impact of therapeutic modality on outcomes following repair of ruptured intracranial aneurysms: an Administrative Data Analysis. J Neurosurgery2010;113:795-801.

Arteriovenous Malformations

Key concepts:

  1. Congenital, abnormal connections of arteries and veins without intervening capillaries
  2. Annual rupture risk of brain AVM’s is 2%-4% per year
  3. Treatment may be surgical, endovascular, or with radiosurgery and is performed to prevent intracranial hemorrhage
  4. In some instances, observant management may be appropriate

General Considerations

AVMs are tangles of congenital abnormal connections between artery and vein with no normal intervening capillary bed. Ninety percent of AVMs are found in the supratentorial space with the remainder found in the brainstem and spine. AVMs can present at any time but are more common in younger patients. Vascular malformations are commonly seen in neurosurgical practice and with modern imaging techniques are increasingly diagnosed in asymptomatic patients being evaluated for headaches or after minor head trauma. Because of this, as well as their risk of rupture and the high morbidity associated with intracranial hemorrhage, it is important that all acute care physicians be aware of their presentation, initial management, and treatment of these lesions.

Epidemiology and Presentation

As many AVMs are asymptomatic, it is difficult to absolutely determine their prevalence. Autopsy data estimate that AVMs occur in less than 1%-4% of the population. A limited number of population-based studies have been conducted to study the natural history and bleeding risks in patients with AVMs. In counseling patients, the following formula has been used to estimate lifetime of intracranial hemorrhage due to AVM rupture:

Lifetime risk (%) = 105 − patient age in years

Hemorrhage is the most common presentation, occurring in over 50% of patients. Because of this, many patients with AVMs are initially evaluated in the emergency department. Seizures and headaches are also frequent, especially with larger lesions. Spinal AVMs may cause back or radicular pain, lower extremity weakness, gait disturbance, or incontinence. Large AVMs with high-volume venous drainage can cause steal phenomena; focal neurologic deficits may result from decreased tissue perfusion of surrounding brain. Increasingly, patients are referred after AVM’s are found incidentally by computed tomography (CT) or magnetic resonance imaging (MRI).

Initial Evaluation & Care

The initial evaluation of AVMs depends upon the patient’s presentation. Many patients who present with rupture are stable, neurologically and medically. The majority of intracranial bleeding from AVM rupture is intraparenchymal. Although there may be an aspect of subarachnoid hemorrhage, it is rare to see isolated subarachnoid hemorrhage in the setting of AVM rupture. In contrast to aneurysmal subarachnoid hemorrhage, where the presence of clot in the subarachnoid space is often immediately devastating, AVM bleeds are less likely to cause death. Also in contrast to aneurysmal subarachnoid hemorrhage, vasospasm is a relatively rare event with AVM associated hemorrhage. This is not to suggest a ruptured AVM is a minor problem—mortality associated with a single hemorrhage is estimated at 10% and morbidity 30%. However, the heterogeneous nature of presenting complaints in patients with AVMs helps explain why the initial approach to patients with AVM’s varies from those with aneurysms, especially in the acute setting.

In ruptured patients, after ensuring the fundamentals of emergency care—airway, breathing, and circulation—are addressed, a complete neurological evaluation should be performed. Neurosurgical consultation is appropriate. Blood pressure should be maintained in the normal range. In any patient with suspected intracranial hemorrhage, a noncontrast head CT should be obtained as soon as possible. Depending upon physician and institutional preferences, CT angiography (CTA) or conventional diagnostic angiography (Figure 36–23) may then be performed to further evaluate the angiographic architecture of the AVM and to guide treatment. MR angiography (MRA) takes longer to obtain than CTA and the angiographic image quality, in our opinion, is suboptimal when compared to CTA; it is not routinely used in the acute setting.

Figure 36–23. Cerebral angiogram demonstrating a frontal arteriovenous malformation filling from the left anterior cerebral artery.

With an initial CT scan that appears consistent with AVM rupture, a standard approach includes conventional angiography. This provides the neuroendovascular team the potential to proceed with AVM embolization during the same procedure as the diagnostic angiogram. Alternatively, for small or deep AVMs, CTA may be appropriate, as these types of lesions are often treated with radiosurgery and the risks of diagnostic angiography may be avoided. In some instances, CTA may not provide a sufficiently clear picture of the AVM and conventional angiography will need to be performed. In this case, it is important to closely monitor renal function due to the multiple contrast dye loads. Adequate hydration with intravenous fluids is important, and in patients with renal insufficiency, bicarbonate infusion and Mucomyst are useful for renal protection. Any patient with intracranial hemorrhage, even those who are neurologically intact, should initially be admitted to the neurosurgical intensive care unit for close monitoring with hourly neurologic examinations.

In patients presenting with seizures or those with large bleeds and mass effect, treatment with antiepileptic medication is appropriate. Phenytoin or Keppra are effective in the acute setting for seizure prophylaxis.

An important concept when discussing brain AVM treatment options and prognosis with patients and the multiple physicians who may care for patients with these lesions is AVM grade. Grading is most often performed using the Spetzler–Martin scale (Table 36–7). Points are assigned based upon size (< 3 cm, 1 point; 3-6 cm, 2 points; or > 6 cm, 3 points), venous drainage (superficial, 0 points; or deep, 1 point), and eloquence (absent, 0 points; present, 1 point) of the surrounding brain, yielding Grades of I-V. The size and venous drainage categories are straightforward; eloquent areas are defined as the sensorimotor, language, and visual cortex; the thalamus and hypothalamus; the internal capsule; the brainstem; the cerebellar peduncles; and the deep cerebellar nuclei. AVM grade correlates with surgical results.

Table 36–7 Spetzler–Martin AVM grading scale.

Treatment

Four options currently exist for the treatment of AVMs. They are endovascular embolization, microsurgical resection, SRS, and observant management. Treatment often employs a combination of these approaches. AVMs should be treated in referral centers with significant experience. There is no treatment algorithm for these complex lesions. Multiple variables intrinsic to the AVM or the patient have been implicated as making certain lesions higher risk for bleeding; most of these are controversial. It is important to evaluate each patient individually, preferably with a multidisciplinary team of vascular neurosurgery, interventional neuroradiology, and radiation oncology. Multiple factors must be considered prior to recommending treatment, including AVM grade and location (for surgical safety), angiographic architecture and presence of reachable arterial pedicles (for endovascular safety), and the ability of the patient to safely tolerate an invasive procedure. Patient preference is also important, especially when considering radiosurgery. Treatment is primarily performed because of intracranial hemorrhage—to prevent an initial or recurrent hemorrhage, or to evacuate the intracranial clot that occurs after AVM rupture. Secondary treatment goals include the relief of mass effect causing headache or seizures.

  1. Endovascular Embolization

The goal of endovascular embolization for AVMs is usually to reduce the size of the nidus and risk of bleeding during microsurgical resection or to reduce the size of the AVM prior to radiosurgery. However, in some cases (10%-20%) complete cure can be achieved with embolization alone. It is important to appropriately counsel patients prior to AVM embolization that if complete occlusion of the AVM cannot be achieved, further treatment with surgery or radiosurgery is necessary as incompletely embolized AVMs may have a higher risk of bleeding.

  1. Microsurgical Resection

As complete obliteration is the only way to cure an AVM, microsurgical resection for small, superficial lesions is the gold standard by which all other treatment modalities are measured. Most neurosurgeons agree that Spetzler–Martin Grade I-III AVMs on the cerebral convexity should be surgically resected. Complication rates associated with these lesions are low, when they are operated on by neurosurgeons with significant experience. Spetzler and Martin retrospectively reported the risk of minor and major neurologic deficit and death in a series of 100 patients with Grade I–V AVMs. For Grade I patients, risk of minor and major deficit was 0%; Grade II AVMs carried a 5% risk of minor deficit and 0% risk of major deficit; Grade III lesions had a 12% risk of minor deficit and 4% risk of major deficit. There were no deaths. Complication rates are higher for Grade IV and V AVMs. Grade IV AVMs carried a 20% risk of minor deficit and 7% risk of major deficit. With Grade V lesions, risks of minor and major deficit were 19% and 12%, respectively. The prospective application of the Spetzler–Martin scale in 120 patients revealed permanent, major neurological deficits in 0% of Grade I-III patients, 21.9% of Grade IV patients, and 16.7% of Grade V patients. The relatively high risk of neurological deficit in Grade IV and V patients make recommending surgery for these lesions a difficult decision. As endovascular technology has improved, however, some of these lesions may be approached endovascularly first, with the goal of attempting to reduce the size of the AVM prior to surgical resection.

  1. Stereotactic Radiosurgery

Radiosurgery is an excellent treatment modality for many AVMs, especially those located in deep locations of the cortex or lesions of the basal ganglia, thalamus, or brainstem that are not easily approachable from a microsurgical or endovascular standpoint. SRS is also indicated for patients with significant medical comorbidities and can be used if an AVM is subtotally resected. In general, SRS works best for AVMs less than 3 cm in size. In patients with AVMs larger than 3 cm, preradiosurgical embolization may be used to reduce the size of the nidus. The complete obliteration rate after SRS is 90% with small AVMs. The primary disadvantage of SRS is the 2-3 years it takes for the AVM to involute. During that period, the patient remains at baseline risk of hemorrhage (~4%) from AVM rupture.

  1. Observant Management

Although this is not the opinion of the vast majority of neurosurgeons and neurointerventionalists, there are occasions when conservative management is indicated. Older patients with comorbid conditions may not benefit from aggressive management. Because of the risk of complications during open or endovascular operations with Grade IV and V AVMs, and because of the reduced rate of complete obliteration after radiosurgery for large lesions, some surgeons advocate conservative treatment for these AVMs.

Postoperative Care

Patients who have undergone endovascular or open operations should be observed postoperatively in an ICU until it is certain they are neurologically stable. Typically, patients are sent to the ward on postoperative day one from microsurgical resection. Postembolization patients are usually discharged home after an overnight stay. Complications after microsurgical resection include the usual postcraniotomy difficulties such as bleeding and seizures. Hydrocephalus can also occur, especially in patients who present with ventricular blood as part of an initial intracranial hemorrhage. Complications to keep in mind after embolization include stroke, renal insufficiency, and groin hematoma. Creatinine and hematocrit should be monitored. Any patient with unstable vital signs or decreasing hematocrit after embolization should be assumed to have a retroperitoneal hematoma, should undergo abdominal CT scanning and be treated aggressively.

After AVM resection, a phenomenon known as normal perfusion pressure breakthrough can occur. As the pathological shunting of blood through the AVM nidus is removed, relative increases in blood flow to the surrounding blood vessels and brain occurs. As these blood vessels are often chronically dysregulated by the relative lack of blood flow caused by the AVM, hemorrhage can result when blood flow returns to normal. One similarly dangerous potential complication after partial embolization is AVM rupture. In general, endovascular neurosurgeons and interventionalists do not embolize more than 1/3 of an AVM at a time. This is because with larger embolizations, dramatic changes in blood flow to the AVM can occur, effectively overwhelming the pathologic vessels remaining, causing hemorrhage. For these reasons, in both instances, patients must be carefully observed postoperatively. Also, it is critical that blood pressure remain in the normotensive range for at least 24 hours posttreatment.

Spinal Arteriovenous Malformations

Spinal AVMs are divided into 4 types: (1) dural arteriovenous fistulas; (2) glomus AVMs; (3) juvenile intradural AVMs; and (4) intradural, extramedullary arteriovenous fistulas. Types 1 and 4 have high blood flow but low pressure; types 2 and 3 have high blood flow, high pressure and are more likely to hemorrhage. Dural arteriovenous fistulas (type 1) are the most common spinal vascular malformation seen. They typically occur near the thoracolumbar junction and consist of a single transdural arterial feeding vessel that directly connects to an intradural arterialized vein. Embolization of the feeding artery, or clip placement across the artery, is curative. Symptoms from spinal AVMs may be due to hemorrhage or venous congestion. Acute or subacute lower extremity neurological deficit, gait difficulties, myelopathy, or loss of bladder or bowel control may result. Initial diagnostic evaluation should include spinal MRI/MRA. Patients with spinal vascular malformations should also undergo cranial imaging to ensure vascular abnormalities of the brain are not present. In patients with negative imaging, spinal angiography is performed.

Summary

Arteriorvenous malformations within the central nervous system are rare. However, with advances in imaging, an increasing number of these lesions are being diagnosed. Arteriorvenous malformations are congenital tangles of direct arterial-venous connections that place patients at risk of hemorrhage of approximately 2%-4% per year. The morbidity and mortality associated with intracranial hemorrhage is high and the most patients with AVMs are young. Because of this, treatment is often recommended to prevent intracranial hemorrhage, or to relieve mass effect causing seizures, headaches, or other neurologic deficits. Treatment options include endovascular embolization, surgical resection, SRS, or a combination of these. With an experienced team of neurosurgeons, interventionalists, and radiation oncologists, treatment can be accomplished with minimal risk. In a minority of cases, observant management may be appropriate.

Spetzler RF, Hamilton MG: The prospective application of a grading system for arteriovenous malformations. Neurosurgery 1994;34:2-7.

SURGICAL MANAGEMENT OF MEDICALLY INTRACTABLE REFRACTORY EPILEPSY

  1. Nicole Bentley Oren Sagher

KEY CONCEPTS

Understand the prevalence, varied etiologies, and broad classification of epilepsy

Describe the various diagnostic modalities for epilepsy

Recognize that focal and lateralizing epilepsy is more amenable to surgery

Recognize that mesial temporal sclerosis is highly epileptogenic with very favorable outcomes following surgical resection

Appreciate palliative and curative surgical treatments for epilepsy

INTRODUCTION

Epilepsy is a disease affecting 1% of the world’s population, and is characterized by a wide range of phenotypes and physical manifestations. The underlying pathology giving rise to epilepsy is equally varied, resulting in a disease that is challenging both to classify and to treat.

Epilepsy is a syndrome of recurrent seizures, the onset of which is due to abnormal neuronal synchronization. The locus of misfiring characterizes a seizure’s semiology, or external signs, and may propagate to surrounding areas. Medical management involves pharmacotherapies and occasionally, alternative therapies, such as dietary restrictions, to reduce seizure frequency. However, one-third of patients are refractory to these therapies. Many studies have demonstrated the efficacy of surgery in cases of failed medical management, and have repeatedly shown that early surgery in appropriate candidates provides the greatest benefit. This chapter will focus on the workup and surgical options in these patients.

CLASSIFICATION

Seizures are classified by their association with alteration in consciousness; “simple” seizures are defined as those that occur in the setting of retained awareness, and “complex” seizures result in loss of awareness. Further classification relies on the locus of seizure onset. If focal, the seizure is considered “partial,” and may be characterized by a stereotypical movement, such as lip smacking or eye deviation. Primary generalized seizures, on the other hand, are considered to affect the entire brain at their onset. Primary generalized epilepsy is usually secondary to a genetic derangement in cell membrane function. A localized group of misfiring neurons may also give rise to auras, or sensory forewarnings, heralding an oncoming seizure. Examples of auras include a taste or light perception, or paresthesias. A partial seizure may spread within the brain—a process known as secondary generalization.

DIAGNOSIS

Several modalities are employed in the diagnosis of epilepsy. These include clinical examination, neurophysiology, imaging, and neuropsychologic assesments.

Clinical Examination & Laboratory Assessment

The clinical symptomatology associated with the epilepsy is a critical component of the diagnosis of epilepsy, and is called its semiology. Symptoms during the time of seizure may provide localizing clues to the region of onset. For example, seizures starting with motor twitching of the upper extremity are likely to be caused by a lesion in the vicinity of the primary motor cortex. Past medical history of febrile seizures or encephalitis are associated with risk of epilepsy. In addition, family history of epilepsy appears to be a strong risk factor in the development of epilepsy. Finally serum metabolic studies should be undertaken in order to rule out potentially reversible entities. Such studies include fasting blood glucose, serum electrolyte panel, complete blood count, and erythrocyte sedimentation rate, renal and hepatic functional assays. In patients where historical data and the clinical examination point to an intoxicating entity, applicable urine and serum toxicology assays should be obtained.

Electrophysiology

The standard modality for recording brain activity is the scalp electro-encephalogram (EEG). EEG recordings are usually obtained both between and during seizures. In some instances, patients undergo long-term monitoring with video EEG where seizure semiology can be assessed together with EEG pattern. Generally, the presence of lateralized or localized seizures is suggestive of a focus that may be amenable to surgical resection. In patients whose seizure localization cannot be demonstrated convincingly by scalp EEG, intracranial EEG electrodes may be placed. These electrodes can either be placed on the surface of the brain (subdural electrodes) or within the substance of the brain (depth electrodes). Subdural electrodes register surface cortical activity, while depth electrodes can provide information related to deep structures such as the hippocampus.

Imaging Studies

MRI is the usual imaging study of choice for evaluating patients with epilepsy. Structural lesions such as tumors, vascular malformations, or dysplastic cortex can be easily identified on MRI. Highly epileptogenic entities such as mesial temporal sclerosis with hippocampal atrophy can also be detected with high resolution MRI scans, as demonstrated in Figure 36–24.

Figure 36–24. Coronal MRI of the brain demonstrating right mesial temporal sclerosis with associated atrophy of the right hippocampus and prominence of the temporal horn of the right lateral ventricle.

Nuclear medicine studies such as positron emission tomography (PET) and single photon emission computed tomography (SPECT) scans can serve as complementary diagnostic tests. These tests are especially useful when the MRI and EEG do not correlate. PET studies demonstrate metabolic activity within the brain, while SPECT studies reflect regional blood flow patterns at the time of tracer injection. SPECT studies performed at the beginning of a seizure often demonstrate increased flow to areas involved in seizure onset. Alternatively, PET studies performed between seizures are more likely to demonstrate hypometabolism within epileptogenic foci. Both studies are useful for patients with focal epilepsy who have normal MRIs, or in whom the site of seizure origin is uncertain.

Magnetoencephalography (MEG) is a functional imaging technique that can accurately provide information on synchronized electrical activity in the brain. MEG detects the magnetic dipole equivalents of electrical current. In addition, it has the advantage of providing a three-dimensional localization of neuronal activity. Finally, there is accumulating data to suggest a high correlation between MEG and intracranial seizure localization.

Neuropsychological Examination

Neuropsychological evaluation is an important component of the epilepsy workup. Patients should undergo a battery of standardized neuropsychological tests assessing verbal and nonverbal intelligence, memory, executive functions, and behavioral functions. These tests often point to subtle deficiencies that accompany the presence of a seizure focus. In addition to these tests, patients may undergo more invasive neuropsychological tests such as the Wada test. The Wada test involves selective injection of a fast-acting barbiturate such as amobarbital into each hemisphere via the carotid artery while memory and language functions are tested. The goal of the Wada test is to assess language and memory dominance. If the Wada test suggests that a significant amount of language function is subserved by the diseased hemisphere, then surgical resection may result in significant deficits.

SURGICAL SELECTION

In general, patients with structural lesions such as tumors and vascular malformations should be managed primarily with surgical resection. Additionally, patients for whom medical therapies have failed to produce an adequate response should be considered for surgery. A consensus definition of medical failure has recently been published, and is defined as failure of two or more antiepilpetic drugs to reduce seizures to a clinically meaningful level, assuming that appropriate doses and regimens have been used. The likelihood of attaining seizure-freedom with pharmacotherapies once two drugs have failed has been shown to be consistently low, less than 5%.

Evidence suggests that early surgical intervention improves outcomes. Delays in surgery may result in encephalopathy, psychosocial abnormalities, learning disabilities, and risk of seizure-related injuries. Additionally, a condition known as SUDEP, or sudden unexpected death in epilepsy, affects a significant number of patients suffering from epilepsy, presumably occurring secondary to cardiorespiratory compromise. With surgical success, patients can be gradually weaned off anticonvulsants, thereby sparing them of the deleterious long-term effects of these medications. However, in patients with epileptogenic foci within the eloquent cortex, the risk of postsurgical deficits must be considered against the probability of rendering the patients seizure-free. Ultimately, the decision to recommend surgery is made by a multidisciplinary team consisting of epileptologists, neurosurgeons, radiologists, neuropsychologists, and social workers.

GOALS OF SURGERY

Surgical procedures can be viewed as either curative or palliative. Curative procedures are designed for patients with seizures convincingly localized to a specific cortical region which is safe to remove. The goal in curative surgery therefore is complete resection of the affected cortex, and procedures include anterior temporal lobectomy, selective amygdalohippocampectomy, neocortical resections, and hemispherectomy. Palliative surgery is employed in situations where a seizure focus is either not identified or cannot be safely removed. For example, patients with congenital syndromic epilepsies such as Lennox–Gastaut syndrome experience life-threatening generalized seizures for which there is no identifiable focus. The goal of palliative surgery therefore is reduction in seizure frequency and severity. Common palliative procedures include placement of a vagus nerve stimulator and corpus callosotomy.

SURGICAL TECHNIQUES

Temporal Lobectomy

The most common cause of drug-resistant epilepsy is mesial temporal sclerosis. Among its several radiographic features, this entity is characterized by hippocampal atrophy on MRI, however, the relationship between the presence and timing of MRI abnormalities and pathophysiology remains poorly understood. While abnormalities on MRI are not an absolute necessity for successful temporal lobectomy, rates of seizure-freedom are lower. Consensus guidelines have recommended that patients who are medically refractory to drug therapy, having failed two or more appropriately prescribed anticonvulsants, should be evaluated by a multidisciplinary epilepsy surgery center.

The surgical technique involves a temporal craniotomy with en bloc resection of 3.5 cm (dominant) or 4 cm (nondominant) of the lateral temporal lobe. En bloc resection of the amygdala, parahippocampal gyrus, and hippocampus is also performed. The long-term seizure-free outcomes are around 60%-80%, with strong trends toward improved quality of life in surgical groups. Though studies have been limited by small sample sizes, there are several well-designed trials that continue to support surgery for appropriate candidates, with a safety profile similar to medically treated groups.

Potential complications include minor visual field deficits and deficits with short-term memory.

Selective Amygdalohippocampectomy

As mentioned earlier, the extent of surgical resection is limited in mesial temporal sclerosis of the language-dominant hemisphere. However, minimally invasive techniques have been developed that spare more of the overlying temporal cortex, minimizing damage to the language pathways. This “selective” procedure involves resection of the amygdala and hippocampus via a temporal craniotomy. The approach to these structures is variable, either through a slit in the temporal lobe cortex or through the sylvian fissure. A comparison of the standard anterior temporal lobectomy to selective amygdalohippocampectomy found that seizure freedom was comparable in the two groups, with 85.2% and 93.1% of patients experiencing seizure-freedom at 3-years, respectively. Potential complications are similar to nonselective lobectomies.

Extratemporal Resections

These procedures are performed in patients with focal epilepsy arising outside the temporal lobe. The frontal lobe is the most common location for such resections. In light of the potential overlap with functional areas, resection is often limited and seizure-free outcome is not as favorable as that seen in temporal lobe epilepsy, averaging approximately 50%. Surgery entails a craniotomy with resection of cortex, at times guided by intraoperative EEG recordings. Complications are contingent upon associated eloquent areas involved.

Hemispherectomy

The hemispherectomy or functional hemispherotomy involves surgical resection or disconnection of one hemisphere from the other. The procedure is used in patients with diffuse unilateral hemispheric seizures that typically result in a neurologically devastated or nonfunctional hemisphere. Hemispherectomy can control seizures in about 70%-90% of patients, with the best outcomes being reported in Sturge–Weber syndrome, Rasmussen’s Encephalitis, and porencephaly. In properly selected cases, function is preserved with associated improvements in cognitive, behavioral, and motor domains.

As originally described, hemispherectomy entailed complete removal of half of the brain. This was fraught with progressive postoperative neurological deficits secondary to deposition of iron over the brain (superficial cerebral hemosiderosis). In light of these problems, a modified functional hemispherectomy has replaced the anatomic hemispherectomy. This procedure involves disconnecting the corpus callosum and the various interhemispheric commissures. The major risks include hemorrhage, damage to functioning cortex, persistent seizures, disseminated intravascular coagulation, and transient decrease in contralateral muscle tone.

Corpus Callosotomy

This procedure is palliative and is utilized in situations where patients have diffuse epilegtonic foci involving both hemispheres. Sectioning the corpus callosum prevents interhemispheric propagation of seizures and is used in both pediatric and adults, with varying extent of sectioning based on patient age and EEG. The procedure typically involves resection of the anterior two-thirds of the corpus callosum, with the option of further resections if seizures persist. Efficacy of the procedure is greatest in drop attacks that cause sudden atonic attacks, with a reported response of 88%. Less responsive seizure types include generalized tonic-clonic, absence, complex partial, and simple partial, with success ranging from 14% to 40%.

Multiple Subpial Transections

This procedure is indicated for patients with seizure foci within functionally important cortex, such as primary motor cortex. The procedure is palliative, with reported efficacy ranging from 33% to 46%. Shallow vertical transections are created across cortical gyri, disrupting the horizontal connections believed to propagate seizures between vertical columns of cells. The procedure is often performed with cortical resections, which has limited the ability to conclusively define outcomes in many studies.

Vagus Nerve Stimulation

The vagus nerve stimulation (VNS) is a relatively low-risk surgical procedure that decreases seizure frequency in patients whose disease is not safely amenable to resection. The exact mechanism by which this technique curbs epilepsy is unknown, but may be due to innervation of the nucleus solitarius which sends fibers to hemispheric regions involved in seizure onset. The stimulator electrode is placed along the vagus nerve in the neck, which is then connected to a generator implanted on the anterior chest wall. In multi-institutional double-blinded, randomized controlled trials, a reduction of seizure frequency of 24%-31% has been reported. A greater response was seen in the pediatric population, with a seizure reduction rate of 50%-90%. Although not considered a curative, intervention, 2% of patients become seizure-free. The main risks of the procedure include injury the vagus nerve, carotid artery, and jugular vein; however, this remains a low-risk intervention in comparison to other surgical modalities.

Stereotactic Radiosurgery

The application of highly focused irradiation to specific brain regions, known as radiosurgery, has been used to treat a wide variety of brain lesions such as tumors and vascular lesions. The role of radiosurgery in seizure control has been fairly limited. However, it has been proposed as a potential treatment of deep-seated seizure foci, such as hypothalamic hamartomas, which are benign malformations in the hypothalamus associated with gelastic epilepsy. Surgical resection in this location is fraught with risk, and it appears that radiosurgery may be used more safely in this subset of patients.

FUTURE DIRECTIONS

Many avenues of surgical treatment for refractory epilepsy are being developed in response to advancements in technology. One example is a closed-loop system implanted into the brain that detects seizure onset and automatically responds by delivering pulsed stimulation to the focus. Preliminary trials have reported seizure reductions of 50%-75%. Stimulation of the anterior nucleus of the thalamus has also been shown to result in seizure reduction, though the mechanism is poorly understood. Stem-cell therapies, gene therapies, and advances in neuronal visualization are also providing tools that increase the understanding and successful treatment of refractory epilepsy.

Anderson WS, Kossoff EH, Bergey GK, Jallo GI: Implantation of a responsive neurostimulator device in patients with refractory epilepsy. Neurosurg Focus 2008;25:E12.

Berg AT, Mathern GW, Bronen RA, Fulbright RK, DiMario F, Testa FM, et al: Frequency, prognosis and surgical treatment of structural abnormalities seen with magnetic resonance imaging in childhood epilepsy. Brain 2009;132:2785-2797.

Engel J, Jr., McDermott MP, Wiebe S, Langfitt JT, Stern JM, Dewar S, et al: Early surgical therapy for drug-resistant temporal lobe epilepsy: a randomized trial. JAMA 2012;307:922-930.

Iida K, et al: Characterizing magnetic spike sources by using magnetoencephalography-guided neuronavigation in epilepsy surgery in pediatric patients. J Neurosurg 2005;102(2 Suppl):S187-S196.

Immonen A, Jutila L, Muraja-Murro A, Mervaala E, Lamusuo S, Kuikka J, et al: Long-term epilepsy surgery outcomes in patients with MRI-negative temporal lobe epilepsy. Epilepsia2010;51(11):2260-2269.

Lee KJ, Shon YM, Cho CB: Long-term outcome of anterior thalamic nucleus stimulation for intractable epilepsy. Stereotactic and functional neurosurgery 2012;90:379-385.

Milby AH, Halpern CH, Baltuch GH: Vagus nerve stimulation in the treatment of refractory epilepsy. Neurotherapeutics 2009;6:228-237.

Sagher O, Thawani JP, Etame AB, Gomez-Hassan DM: Seizure outcomes and mesial resection volumes following selective. Neurosurg Focus 2012 Mar;32(3):E8.

Schachter SC, Guttag J, Schiff SJ, Schomer DL: Advances in the application of technology to epilepsy: the CIMIT/NIO Epilepsy Innovation Summit. Epilepsy Behav: E&B 2009;16:3-46.

Schramm J, Kuczaty S, Sassen R, Elger CE, von Lehe M: Pediatric functional hemispherectomy: outcome in 92 patients. Acta Neurochir (Wien) 2012;154:2017-2028.

Sunaga S, Shimizu H, Sugano H: Long-term follow-up of seizure outcomes after corpus callosotomy. Seizure 2009;18:124-128.

SURGICAL MANAGEMENT OF PAIN

Arnold Etame Parag G. Patil

The subjective, emotional, and physical components of pain make its management complex. Hence, pain management is an interdisciplinary endeavor encompassing medical, surgical, and psychological treatment modalities. Surgery for pain should be reserved for patients who have been unresponsive to therapies directed toward the inciting processes and to oral pain medications.

Pain may be classified as nociceptive or neuropathic. Nociceptive pain results from tissue injury. Common characteristics include constant aching or throbbing and responsiveness to opiate medications. Neuropathic pain is initiated or caused by a primary lesion or dysfunction in the nervous system. Common characteristics include burning, allodynia, and paresthesias. Neuropathic pain responds poorly to opiate medications. Pain surgeons should be familiar with these concepts to assess medical intractability.

The aim of surgery is to interrupt pain signaling pathways. Ablative surgical techniques involve physical interruption through the destruction of neural tissue. Nonablative procedures involve functional interruption through the modulation of pain transduction mechanisms.

ABLATIVE PROCEDURES TO INTERRUPT AFFERENT PAIN PATHWAYS

Neurosurgical procedures to physically interrupt pain signaling have been directed toward the nerves (neurectomy), spinal roots (rhizotomy), dorsal root ganglia (ganglionectomy), dorsal root entry zone (DREZ lesioning), spinal cord (cordotomy, myelotomy), and cerebral cortex (cingulotomy). Ablative surgery for pain is most often utilized in the treatment of cancer-related, nociceptive pain, as long-term analgesia is less commonly observed for these procedures.

Neurectomy

Neurectomy involves cutting an injured nerve or the nerve to a painful region. Target nerves are identified on the basis of local anesthetic blockade. Denervation of joints, distal sensory nerves, and neuroma surgery are examples of peripheral neurectomy. Neurectomy is not typically utilized for cancer-related pain because of the changing pain distribution with tumor growth. Reported success rates for pain control with neurectomy vary widely between 40% and 90%.

Rhizotomy & Ganglionectomy

Rhizotomy and ganglionectomy target the dorsal sensory rootlets or ganglia, respectively. Spinal cord segments are identified through paraspinal local anesthetic blockade, with placebo controls. The procedures are utilized most commonly for cancer-related regional pain or occipital neuralgia. These procedures are rarely utilized in the treatment of extremity pain because of functional impairment resulting from the loss of proprioception. Successful longer term pain control has been reported in 40% to 70% of patients.

Dorsal Root Entry Zone Lesioning

Dorsal root entry zone surgery targets the superficial dorsal horn region, where sensory fibers enter the spinal cord. Levels including and flanking the region of interest, as defined by imaging studies or pain distribution, are typically ablated. A knife or radiofrequency heating is used to make the lesion. Dorsal root entry zone surgery is most effective in the treatment of neuropathic pain following nerve root avulsion and at-level spinal cord injury pain. Risks of surgery include injury to descending motor pathways and decreased sensory function in the territory of the ablated region. In carefully selected patients, rates of successful pain control range from 70% to 90%.

Cordotomy & Myelotomy

Cordotomy is a spinal procedure to interrupt pain transmission along the lateral spinothalamic tract. Cordotomy is most commonly performed in patients with intractable, unilateral, nociceptive cancer pain at the level of the chest or below. The procedure may be performed either with a knife or with radiofrequency heating. Risks of surgery include lower extremity weakness, ataxia, and respiratory or urinary dysfunction. These risks are significantly increased when cordotomy is performed bilaterally.

Midline myelotomy involves destruction of the mesial dorsal columns at a single spinal level to treat midline, bilateral, or visceral pain. The procedure typically preserves dorsal column and spinothalamic signal transmission. Risks include transient lower extremity paresthesias and weakness.

Rates of successful pain control with cordotomy and myelotomy are initially high (> 80%) but decline over time (40% at 2 years).

Cingulotomy

Unlike procedures directed along pathways of pain neurotransmission, cingulotomy is directed toward alteration of the experience of pain. Because the cingulate gyrus is part of the limbic system, radiofrequency ablation of the anterior cingulate gyrus reduces the affective, unpleasant aspects of pain, particularly in patients with obsessive and affective components to their pain. Cingulotomy is performed in relatively few centers and only in carefully selected patients. Following cingulotomy for intractable, cancer-related pain, over 50% of patients have been reported to have moderate to complete pain relief.

PROCEDURES TO MODULATE AFFERENT PAIN PATHWAYS

Intrathecal Analgesic Delivery Pumps

When compared to oral narcotics, intrathecal administration of morphine provides more potent analgesia with reduced side effects, such as nausea, constipation, and sedation. To benefit from intrathecal delivery, patients should have significant reduction in pain level with oral opiates, limited by intolerable side effects.

Analgesics such as morphine are delivered through a catheter placed into the cerebrospinal fluid of the spinal canal. The catheter tubing is then connected to an external or surgically implanted pump. External pumps are utilized in patients with cancer-related pain and expected survival of less than 3 months. Principal complications of intrathecal drug delivery include the side-effects of the medication, mechanical failure of the system, and infection.

Intrathecal drug delivery may be effective in either nociceptive or neuropathic pain syndromes. However, as with oral opiate administration, long-term tolerance to medications can develop. Hence, these devices are most beneficial for patients with cancer-related pain syndromes and limited life expectancy.

Peripheral Nerve & Spinal Cord Stimulation

Peripheral nerve and spinal cord stimulators deliver pulses of electricity to injured nerves or the dorsal columns of the spinal cord, respectively. According to the gate theory of pain, such stimulation blocks the flow of pain signals from the periphery to the brain.

Peripheral nerve stimulation is most effective in neuropathic peripheral nerve syndromes such as occipital neuralgia and complex regional pain syndrome. More recently, peripheral stimulation has been used to treat headaches and fibromyalgia. Spinal cord stimulation is most effective in patients with lumbosacral radiculopathy due to scar tissue formation following back surgery as well as in patients with complex regional pain syndrome.

Patient candidates typically undergo an initial trial with a temporary electrode applied to the spinal cord or nerve. After the 1-week trial, permanent placement is performed if benefit is demonstrated. Benefit is measured as a reduction in pain as well as an increase in daily activities. Complications of the therapy are most commonly stimulating lead migration, breakage, and infection.

Deep Brain Stimulation

Deep brain stimulation involves the precise surgical placement of electrodes into the deep nuclei of the brain. Common targets of deep brain stimulation for pain include the thalamus, which is the sensory relay of the brain, and the periaqueductal gray region, which results in the upregulation of endogenous opiates. Once electrodes are implanted, trial stimulation is performed for 1-2 weeks. A successful trial is characterized by the experience of pain-relieving and tolerable paresthesias in the treated region during thalamic stimulation, a sense of warmth and ocular movement during periaqueductal stimulation, a poststimulatory pain-relieving effect, and the absence of an analgesic effect during sham stimulation. Following the trial, a pulse generator is connected to the wires and implanted in the chest. Long-term results in patients with a successful trial are variable, ranging from 19% to 79%.

Motor Cortex Stimulation

Electrical stimulation of the region of the motor cortex results in analgesia in neuropathic pain syndromes such as hemibody poststroke pain and trigeminal deafferentation pain. The mechanism of motor cortex stimulation is unknown.

The surgical procedure involves placement of a stimulating electrode under the skull in the region of the motor cortex. Patients undergo a trial of stimulation. Stimulus intensity is typically set to 80% of the level needed to produce motor cortical responses. Following a successful trial, the electrodes are connected to an implantable pulse generator in the chest. Motor cortex stimulation has a success rate of 70% in facial pain syndromes and of 50% in central neuropathic pain.

SURGICAL MANAGEMENT OF MOVEMENT DISORDERS

Arnold Etame Parag G. Patil

Neurosurgical procedures for movement disorders have evolved considerably in recent years. Formerly popular stereotactic tissue-destructive procedures, such as pallidotomy and thalamotomy, have been superseded by nonlesional deep-brain stimulation (DBS). Careful, prospective and well-controlled studies have demonstrated significant benefits of DBS in the treatment of Parkinson disease, essential tremor (ET), and dystonia.

PARKINSON DISEASE

Clinical Considerations & Pathophysiology

James Parkinson was the first to describe the “shaking palsy” in 1817. The clinical signs of Parkinson disease (PD) are tremor, bradykinesia (slowness of movement), rigidity (increased muscle tone), and postural instability. The tremor of PD occurs at rest, has a “pill-rolling” character, and typically decreases with voluntary movement. The rigidity of PD has a “cogwheel” ratchet-like quality during passive movement. Postural instability results from a loss of reflexes, leading to impaired balance. Other signs of PD include a shuffling gait, decreased voice volume, slowed reaction time, and dementia. A popular scale for the measurement of Parkinsonism is the Unified Parkinson Disease Rating Scale (UPDRS).

PD is accompanied by a loss of dopaminergic neurons in the substantia nigra pars compacta. According to a well-accepted model of basal-ganglia function, the loss of dopamine results in activation of the subthalamic nucleus and the globus pallidus pars interna (GPi). The GPi inhibits motor regions of the thalamus, resulting in decreased cortical excitation and the symptoms of PD. The central role of the GPi and subthalamic nucleus in this scheme provides the impetus for the surgical therapies for PD.

Idiopathic PD must be distinguished from other Parkinsonian syndromes that have similar signs and symptoms. These syndromes include multiple system atrophy, progressive supranuclear palsy, corticobasal degeneration, and dementia with Lewy bodies. There are no laboratory or blood tests that help in the diagnosis of PD. CT and MRI studies of patients with PD are typically normal. For each patient, the diagnosis is based entirely on the history and physical examination as well as responsiveness to medication. As a result, only 75% of patients with a clinical diagnosis of PD are confirmed at autopsy.

Medical Management

Medical management strategies for the treatment of PD center upon manipulation of the dopaminergic system. l-dopa, which was introduced in 1967, crosses the blood-brain barrier and is converted by dopaminergic neurons into dopamine. l-dopa is often compounded with carbidopa, an inhibitor of dopamine metabolism in the bloodstream, to increase the efficiency of l-dopa delivery to the brain. Other medications that are useful in the treatment of PD include inhibitors of the COMT and MAO-B enzymes, which metabolize dopamine, as well as direct agonists of dopamine receptors in the brain. Patients with non-PD, atypical Parkinsonian syndromes do not respond well to l-dopa therapy.

Over 5-10 years, PD patients develop several troublesome side effects of l-dopa. Dyskinesias are involuntary writhing movements of the face and extremities that occur at peak dopamine levels. In addition, after chronic therapy, patients may develop on-off fluctuations in which their Parkinsonian symptoms oscillate in an unpredictable manner. Finally, patients may develop involuntary freezing during movement. The presence of such side effects of l-dopa should prompt surgical evaluation.

Surgical Management

Lesional stereotactic surgery for PD has been performed since the 1950s. Principal targets for ablation include the thalamus (thalamotomy) and the GPi (pallidotomy). With the introduction of l-dopa, these lesional surgical therapies declined. However, with the appearance of l-dopa side effects in the 1980s and with the development of DBS techniques in the 1990s, surgery for PD has increased significantly. At present, DBS of the subthalamic nucleus or GPi is favored over lesional surgery because of its heightened safety and reversibility. Since FDA approval in 1997, over 10,000 patients with PD have been treated with DBS.

Surgical indications for the treatment of PD are well established. Guidelines defined by the Core Assessment Program for Surgical Interventional Therapies in PD (CAPSIT-PD) include the following:

  • A diagnosis of idiopathic PD for a period of 5 years
  • Exclusion by history and MRI of atypical Parkinsonism
  • Dopaminergic responsiveness (33% reduction in UPDRS motor score with l-dopa)
  • No significant cognitive deterioration or depression

The goal of surgery for PD is an improvement in motor symptoms. DBS results in significant improvements in tremor, rigidity, bradykinesia, postural stability, freezing, and gait, compared to the off-l-dopa state. DBS is not expected to provide improvement to patients with PD beyond their best on-l-dopa state. However, as doses of l-dopa are typically lowered after DBS surgery, DBS provides relief from the dyskinesias and on-off fluctuations associated with chronic l-dopa therapy.

A recent study has determined that DBS results in a significant improvement in quality of life for patients with PD. Compared to medication alone, DBS provides increased ability to perform activities of daily living, improved emotional well-being, decreased stigma of disease, and reduced bodily discomfort. Benefits are likely to be reduced, however, in patients over 70 or with significant cognitive deficits in whom motor improvements alone are unlikely to alter quality of life significantly.

ESSENTIAL TREMOR

Clinical Considerations & Pathophysiology

ET is the most common movement disorder, affecting an estimated 2% to 4% of the population. ET can be present in adolescence but most often appears in middle age or later life and is slowly progressive. There is a strong genetic component, with 25% to 60% of patients reporting a family history of tremor, typically with an autosomal dominant pattern of inheritance.

The tremor of ET occurs during maintenance of posture against gravity and with action. ET can occur in any part of the body but is most commonly observed in the hand (90%-100%), head (40%-60%), and voice (25%-35%). The tremor of ET is distinguished from rest tremors, which occur when a limb is fully supported against gravity, and from intention tremors, which occur during visually guided movement as the limb approaches the target. However, in severe cases, patients with ET may experience either rest or intention tremor in addition to action tremor.

The pathophysiology of ET is not well understood, though cerebellar function appears to be involved. In addition to tremor, patients with ET may have mildly ataxic or dysmetric gait, oculomotor deficits, and disordered eye-hand movements, reminiscent of cerebellar dysfunction. In addition, PET studies have demonstrated increased cerebellar activity in patients with ET. It is thought that disruption to olivocerebellar rhythmicity may be central to the development of ET.

ET must be differentiated from other tremor disorders. Disease processes resulting in action tremor may include PD, enhanced physiological tremor, dystonia, and Wilson disease. Other tremor disorders may include cerebellar (intention) tremors, Holmes (rubral) tremor, toxic/metabolic disorders, and psychogenic disorders.

Medical Management

In many cases, ET begins late in life, progresses slowly, and is neither physically disabling nor psychologically burdensome. Some patients may experience reduced tremor by restricting or eliminating caffeine from the diet or by wearing small weights about their wrists. Some patients may experience reduced tremor with moderate alcohol consumption. However, alcohol is not typically recommended as a treatment because of risks of resulting chemical dependence in susceptible individuals.

For patients with disabling tremor, first-line therapies include beta-blockers, such as propranolol, and the antiepileptic primidone. Approximately 50%-70% of patients obtain benefit from beta-blockade. Primidone is a medication related to phenobarbital, with similar efficacy to beta-blockade in the treatment of ET. In some patients, the two therapies may be combined for an additive effect. Additional pharmacological agents for ET include gabapentin, topiramate, and long-acting benzodiazepines, such as clonazepam. Finally, a subset of patients with ET may be treated by local botulinum toxin injection.

Surgical Management

As in the treatment of PD, lesional surgery for ET has been largely superseded by DBS. The targets of ablation and DBS are the same, the ventralis intermedius nucleus of the thalamus. The ventralis intermedius nucleus receives inputs from the deep cerebellar nuclei, including the dentate nucleus, which may account for its importance in the treatment of ET. A unilateral procedure is favored for patients with disabling unilateral extremity tremor, while a bilateral procedure may be required to control bilateral or axial tremors.

Thalamotomy of the ventralis intermedius nucleus is highly effective in the treatment of ET, with over 80% of patients experiencing effective long-term tremor suppression. Complications of thalamotomy occur in some 25% of patients and primarily include hemorrhage, weakness, dysarthria, and ataxia.

Ventralis intermedius nucleus DBS has largely replaced thalamotomy in the surgical treatment of ET. In a prospective, randomized study comparing DBS to thalamotomy, both therapies achieve similar tremor control. However, DBS results in fewer adverse effects. DBS may have more favorable effects on patient functional status, including activities of daily living. Complications of thalamotomy and DBS are more pronounced in patients following a bilateral procedure.

PRIMARY DYSTONIA

Clinical Considerations & Pathophysiology

Dystonia is the sustained cocontraction of opposing muscle groups. Patients with dystonia exhibit abnormal and awkward postures, engage in repetitive movements, and often experience significant pain. Dystonia may affect muscles throughout the body (eg, generalized dystonia) or muscles in a region (eg, torticollis), or it may have a specific focus (eg, blepharospasm). Like tremor, dystonia may be an isolated finding or a manifestation of a more generalized neurological condition.

The pathophysiology of dystonia is not known. Some cases of dystonia have been shown to result from dopamine deficiency or disordered function of dopamine receptors in the basal ganglia. One model suggests that decreased or dysregulated activity in GPi results in disinhibition of motor cortical areas.

Dystonia often occurs as an idiopathic condition, without a clear etiology. Alternatively, dystonia may occur as a secondary condition, resulting from birth injury, stroke, drug toxicity, or a hereditary degenerative neurological condition. Recently, over a dozen hereditary forms of previously idiopathic dystonia have been identified, including mutations of the DYT1 gene on chromosome 9. The distinction between primary and secondary dystonia is important, as secondary dystonias respond less well to surgical interventions.

Medical Management

Primary dystonia may be treated with anticholinergic drugs such as trihexyphenidyl, benzodiazepine muscle relaxants such as valium, or the injection of botulinum toxin into affected muscle groups. In addition, some dopamine- blocking medications have been utilized in the treatment of dystonia, although use of such medications also may worsen some forms of dystonia. Physical therapy is also an important component of dystonia treatment to prevent the formation of fixed muscle contractures.

Surgical Management

Patients who fail to respond to oral medications and who fail to achieve adequate relief with botulinum toxin injections should be referred for surgical management. Dystonia has been treated with either ablation or stimulation of the GPi. Pallidotomy improves dystonia by 60% to 70% when measured by standard rating scales. Bilateral pallidal DBS is also highly effective in the treatment of primary dystonia. By comparison to sham stimulation, DBS significantly improved motor symptoms, pain, and quality of life. The most common side effect of pallidal DBS is dysarthria.

TECHNIQUES OF STEREOTACTIC NEUROSURGERY

Stereotactic neurosurgery involves the ablation of tissue or the placement of electrodes deep into the brain. Both the clinical efficacy and risks of surgery depend on submillimeter accuracy, requiring specialized techniques. Patients are typically placed in a stereotactic frame. This frame is affixed to the skull, under local anesthetic. The patient then undergoes an MRI or CT scan. Performance of a scan while in the frame allows a precise coordinate system to be defined within the brain.

In the operating room, a small hole is drilled into the skull, allowing the introduction of microelectrodes along a trajectory leading to the target. The microelectrodes record extracellular neuronal activity. Each region of the brain has a specific electrophysiological signature. Depending on the surgery, the patient may be examined for motor or sensory responsiveness of cellular activity.

Once electrophysiology has confirmed the target for surgery, DBS electrodes are placed into the same location along the same trajectory. With the electrode in position, stimulation is applied and the patient is examined for undesirable clinical effects. Once a desirable effect is confirmed, the electrodes are secured. As a second stage, the electrodes are tunneled under the skin to an implantable stimulation generator placed under the skin, just below the clavicle. This generator may be precisely tuned to the requirements of each patient.

In lesional surgery, a radiofrequency or cryoprobe is placed into the location following microelectrode recording, and a temporary lesion is created. When the absence of undesired effects is confirmed, a permanent lesion is created.

INTERVERTEBRAL DISK DISEASE

  1. Andrew Wilkinson Khoi Than Paul Park

General Considerations

The spinal column is composed of 33 longitudinally stacked bone segments called vertebrae: 7 cervical, 12 thoracic, 5 lumbar, 5 sacral (fused), and 2-4 coccygeal (fused). A typical vertebra is composed of a rounded body anteriorly and a protective boney arch posteriorly, which together form a canal through which the spinal cord passes. Vertebral bodies articulate anteriorly via intervertebral discs and posteriorly via synovial joints formed by the articular facets of adjoining vertebrae. They are also connected by several ligaments: the anterior and posterior longitudinal ligaments which traverse the vertebral bodies from top to bottom, the supraspinous and interspinous ligaments which run between the posterior projecting spinous processes of each vertebra, and the ligamentum flavum which connects the lamina (part of the posterior arch) at each vertebral level.

In a normal adult, the spinal cord extends from the cranio-cervical junction to the lumbar level, where it tapers and typically ends at L1-2 as the conus medullaris. Eight sets of nerve roots exit the spine in the cervical region, though there are only seven cervical vertebrae. The C1 nerve root exits above the C1 vertebra, while the C2 nerve root exits between the C1 and C2 vertebrae. The C8 nerve root exits between the C7 and T1 vertebral bodies, and the T1 nerve root exits between the T1 and T2 vertebral bodies. This relationship continues downward to the level of the sacrum. Thus, the nerve root that emerges at the L5-S1 level is the L5 nerve root. In the case of lumbar disk herniation, the affected nerve root is typically the root that is passing by to exit at the next level. As an example, a disk herniation at the L4-5 level would typically compress the L5 nerve root. Conversely, in the cervical spine the nerve that is affected is at the level of the disk herniation. A C5-6 disk herniation would therefore impact the C6 nerve root.

Intervertebral disks act as pads separating the vertebral bodies of the spine. They also function as shock absorbers, helping to cushion and distribute downward forces on the spine. Additionally, intervertebral disks allow a limited amount of movement to occur between different spinal levels so that the spine may bend and rotate. The intervertebral disk is composed of a gel-like, elastic fibrocartilaginous central nucleus (the nucleus pulposis) surrounded by a fibrous outer ring (the annulus fibrosis) composed of 15-25 concentric layers of parallel fibers. Vertebral body end-plates less than 1 mm thick and composed of hyaline cartilage form an interface between the bone of the vertebral bodies and the disk, sandwiching the nucleus pulposis superiorly and inferiorly.

Intervertebral disks degenerate over time. Coupled with degeneration of the facets, this process is termed spondylosis. At birth, the nucleus pulposis contains 80% water. As people age, however, disks gradually lose their water and elasticity, becoming less gel-like. The process of disk degeneration is common and may even be “normal” as people age. About 20% of teenagers have signs of mild disk degeneration, whereas by age 70, approximately 60% of disks are severely degenerated. Degenerated disks do not distribute load in the same way as healthy, well-hydrated disks. They do not maintain their height under load-bearing conditions and, as a consequence, more load is placed on vertebral bodies and adjacent facet joints. This promotes the formation of osteophytes. If osteophytes form in the spinal canal, within the neural foramina, or in the lateral recess, neurologic compression may develop over time. With loss of disk height, the tensional forces on the ligamentum flavum are reduced, causing the ligmentum to remodel, thicken, and bulge into the canal (ligamentous hypertrophy). Degeneration can also lead to spondylolisthesis or subluxation of one vertebral body over another. Degenerative disk disease occurs at all levels of the spine; however, because the lumbar spine and cervical spine have greater mobility, pathology is more common in these regions.

CERVICAL SPINE

General Considerations

Degeneration of the cervical spine is a process leading to bulging of intervertebral disks, hypertrophy of facet joints, and osteophyte formation. Chronic degeneration initially manifests as neck pain. As osteophytes enlarge and ligamentous hypertrophy progresses, neurologic symptoms may develop. Compression of nerve roots leads to radiculopathy, while compression of the spinal cord itself leads to myelopathy. Alternatively, disk degeneration may occur acutely when the nucleus pulposis of a disk is extruded through a tear in the annulus. If an acute disk herniation occurs centrally, the spinal cord may become compressed, causing severe neurologic injury. This may result in paraplegia or quadriplegia, depending on the level and severity of the herniation. More commonly, however, disk rupture results in compression of a nerve root by the extruded disk fragment as well as local inflammatory changes, causing radiculopathy. This is typically manifest as pain radiating down the arm, sensory disturbance and, sometimes, weakness in the distribution of the involved nerve root.

Clinical Findings

  1. Symptoms & Signs

Degenerative disease of the cervical spine often presents with a history of neck pain which may either be abrupt, in the case of disk rupture, or slowly progressive. There is often a loss of cervical lordosis (the normal, backward, C-shaped curvature of the neck), which may be related to muscle spasm or to deformity from chronic degeneration. In addition to neck pain, which often abates over time, compression of a single nerve root by an osteophyte or disk fragment (radiculopathy) often causes an aching pain along the medial border of the scapula on the side of the lesion. This scapular pain tends to be longer lasting than the neck pain. The characteristic finding in radiculopathy is a sharp, burning pain that radiates down the arm, following the distribution of the involved spinal nerve. This pain may be exacerbated when the patient tilts the head toward the side of the pain, crowding the neural foramina on the affected side (Spurling maneuver). Indeed, the patient may habitually tilt the head to the opposite side to reduce the pain. Hyperextension of the neck (with or without compression of the head) may worsen the pain. Sensory disturbances (parasthesias, numbness, or decreased sensation) tend to occur in the terminal distribution of the involved dermatome, that is, in the fingers rather than the proximal arm. Hypersensitivity of the skin in the distal distribution of the dermatome is also common. A decrease or loss of deep tendon reflexes is a frequent and early finding in radiculopathy from a herniated disk or a compressive cervical osteophyte. Weakness from radiculopathy occurs in muscles innervated by one spinal nerve (but by more than one peripheral nerve); that is, it is myotome-based. Thus, weakness from radiculopathy is often partial or incomplete, since nearly all muscles are innervated by more than one spinal nerve. Profound weakness, atrophy, and muscle fasciculations are rare in radiculopathy, except in very long-standing cases. The presence of these findings should generate suspicion of a peripheral nerve lesion.

C5 radiculopathy (typically resulting from pathology at the C4-5 level) involves pain radiating into the shoulder, with sensory disturbances crossing over the top of the shoulder and extending to the mid-portion of the upper arm (following the distribution of the C5 dermatome). Patients may exhibit weakness of shoulder abduction (deltoids) and forearm flexion. The biceps reflex may be attenuated. C6 radiculopathy (as from a C5-6 herniated disk) typically involves pain radiating from the neck into the lateral aspect of the arm, with sensory disturbance in the dorsum of the hand and, in particular, the thumb. Patients may present with weakness of forearm flexion (biceps). The biceps reflex as well as the bracheoradialis reflex may be attenuated. C7 nerve root compression (from C6-7 pathology) often involves pain radiating from the neck into the back of the shoulder, the triceps and the dorsolateral surface of the forearm. Sensory disturbance typically involves the index and middle fingers. Weakness of forearm extension (triceps) is generally noticed in a delayed fashion, perhaps because day-to-day extension of the forearm occurs with the assistance of gravity. The triceps reflex is often attenuated.

In advanced cases of degenerative disk disease of the cervical spine, signs of myelopathy may develop, including hyperreflexia, spasticity leading to gait disturbance, and sensory disturbance in the upper and lower extremities. Patients with myelopathy from cervical stenosis often complain of difficulty manipulating objects with their hands (eg, problems buttoning their shirt).

  1. Diagnostic Studies

Plain films are useful in determining the degree of degenerative change present in the cervical spine. In cases of cervical deformity, plain films are used for evaluating the alignment of the cervical spine. Flexion-extension plain films are important when there is a question of instability of the cervical spine (eg, when a patient is having positional symptoms). Likewise, computed tomography (CT) scans may offer detailed views of the bony anatomy and are often useful for preoperative planning, especially in cases of severe deformity. CT also offers good resolution of boney anatomy when bone spurs are suspected as the cause of neural compression. The major disadvantage of CT scanning is the lack of resolution of soft tissue structures; it is difficult to detect compression from a herniated disk on a normal CT scan. CT myelography solves the problem of visualizing soft tissue compressive lesions. However, its main disadvantage is its invasiveness: Puncture of the thecal sac (required for dye injection) carries a small risk of neurologic injury. In the era of magnetic resonance imaging (MRI), CT myelography is often reserved for cases where the spine has been previously instrumented, which may cause significant artifact on MR imaging, or if the patient has other metallic devices that preclude magnetic scanning (eg, a cardiac pacemaker).

MR imaging allows the resolution of neural structures in a noninvasive manner and has become the most common imaging method for evaluating potentially compressive pathology of the cervical spine. MRI can detect soft tissue disk herniation and nerve root compression. It is also useful in detecting chronic or acute changes in the spinal cord that may be associated with myelopathy. Findings on MRI should be carefully correlated with clinical findings, as false positives are frequently generated. MR imaging reveals degenerative disk disease of the cervical spine in 25% of asymptomatic people less than 40 years of age and in 60% of people over 40 (Figure 36–25).

Figure 36–25. Sagittal T2-weighted MRI of cervical spine showing multilevel disk protrusions causing central canal stenosis and cord signal change.

Electrodiagnostic studies, particularly EMG, may be useful in diagnosing radiculopathy. Nerve conduction studies alone are of little value in identifying radiculopathy and are generally normal, even with severe compression of a nerve root. EMG, on the other hand, is more sensitive. Classic EMG findings in radiculopathy are fibrillations at rest in muscles supplied by a single nerve root (ie, a myotome) along with denervation in the corresponding paraspinal muscles. Unfortunately, EMG will not reliably detect fibrillations in muscles until at least 3-4 weeks following the onset of radiculopathy. This may lead to false negative studies if the test is performed too soon. Even when performed after an appropriate waiting period, EMG findings may be normal in upward of 50% of cases of spinal nerve compression in patients with radicular symptoms, but no signs of weakness, numbness, or decreased reflexes.

Differential Diagnosis

Neck pain associated with a history of malignancy, unexplained weight loss, pain unrelieved by bed rest, or age more than 50 with cancer risk factors should raise suspicion of a metastatic tumor invading the cervical spine. Similarly, infectious etiologies such as diskitis, osteomyelitis, or abscess should be considered when there is a history of fever, immunosupression, or recent infection. Peripheral nerve entrapment syndromes such as carpal tunnel syndrome or ulnar nerve compression may mimic cervical radiculopathy. In general, severe weakness and muscle atrophy is suggestive of a peripheral nerve lesion, while early loss of a reflex (biceps, triceps) suggests radiculopathy. Other conditions that may mimic cervical degenerative disk disease include myocardial infarction, idiopathic brachial plexitis (Parsonage Turner syndrome), or inflammatory conditions such as ankylosing spondylitis or sarcoidosis. Local conditions affecting the shoulder (rotator cuff tears, acromial bursitis, etc) must also be ruled out.

Treatment & Prognosis

Most conditions that cause pain in the cervical spine (such as exacerbations of degenerative arthritis, muscle spasm, or minor trauma) are self-limiting and ultimately do not require operation. Acute neck pain may be treated with gentle exercise or a mobilization program, moist heat, or a soft collar to help muscle relaxation. Anti-inflammatory medications are also useful in this regard. For persistent neck pain, intermittent traction is sometimes helpful, either through physical therapy or with a home traction kit. Roughly 80%-90% of patients improve with medical management alone, though many continue to have mild symptoms they ultimately learn to manage.

Neck pain itself responds poorly to operative management. Even in cases of radiculopathy where imaging reveals a clear-cut disk herniation compressing a nerve root, surgical management is most likely to improve only arm pain rather than neck pain. Surgical management of cervical degenerative disk disease should be reserved for cases failing medical management and where there is neurologic compression (leading to either myelopathy or radiculopathy). Operative management of the cervical spine involves decompression of the spinal cord or nerve roots, with or without fusion. The cervical spine may be approached either anteriorly or posteriorly. The choice depends on many factors, including the age of the patient, the number of levels involved, whether the compressive lesion is predominantly anterior or posterior, and any concurrent deformity of the cervical spine. Both anterior and posterior approaches may be used to decompress nerve roots and/or the spinal cord. For complicated cases involving extensive degenerative change, particularly with severe deformity, a combined anterior/posterior approach may be employed.

Disk herniations and osteophytes may be addressed anteriorly, either by removing just the disk (anterior cervical discectomy, with or without fusion) or by drilling away the vertebral body (a procedure known as corpectomy). Posterior cervical laminectomy is useful for decompression of multiple levels, as in the case of multilevel cervical stenosis secondary to ligamentous hypertrophy. Because there is a risk of subsequent deformity (progressive kyphosis related to loss of the posterior tension band following operation), some patients who are approached posteriorly may need to be fused. The decision to fuse should be made on a case-by-case basis. Artificial disk replacement (arthroplasty) in lieu of fusion has been shown to maintain segmental mobility and appears to be a viable alternative to fusion, though long-term results are not yet available. Posterior keyhole foraminotomy is ideally suited for soft disk herniations that occur laterally (it cannot be used for central disk bulges) and may be done in a minimally invasive fashion using tubular retractors.

In the case of cervical radiculopathy, symptoms improve in approximately 80% of patients following operative management. Where surgical decompression is performed for myelopathy, neurologic improvement occurs in approximately 70% of cases.

THORACIC DISK DISEASE

Thoracic disk herniations are rare, with an incidence between 0.25% and 0.75% of all disk herniations. The majority of thoracic disk herniations occur below the level of the mid-thoracic spine. Often there is a delay in diagnosis because of poorly defined symptoms and the lack of objective findings on physical exam. If the disk herniation is secondary to trauma and results in severe cord compression, paralysis may be the result. If the disk herniation is secondary to degenerative changes, the cord compression occurs more slowly and is associated with a variety of presentations.

Patients may present with symptoms of axial pain, radiculopathy, myelopathy, or some combination of the three. The axial pain may be described as dull, aching, burning, stabbing, or cramping. Load bearing, activity, or valsalva will often exacerbate the pain. Radicular symptoms generally present in the appropriate dermatomal band. Myelopathy can present as paraparesis, but more often presents with a vague history of lower extremity weakness, heaviness, stiffness, or numbness. Bowel and bladder complaints can occur.

Treatment is surgical and is directed at alleviating pain or preventing progression of a neurologic deficit. There are a variety of surgical options including laminectomy for stenosis as well as a variety of approaches (thoracotomy, costotransversectomy, lateral extracavitary, transpedicular) for pathology that occurs in the anterior spine, such as a disk herniation. In cases of a thoracic disk herniation, a strictly dorsal midline approach (laminectomy) offers poor exposure of the disk and has a high risk of neurologic injury.

LUMBAR SPINE

General Considerations

One must understand the anatomy of the lumbosacral roots to appreciate the clinical syndromes associated with a displaced lumbar intervertebral disk. An extruded lumbar intervertebral disk can lead to loss of reflexes (ankle jerk, patellar reflex), motor loss, sensory loss, and pain in a dermatomal distribution. A central disk herniation can lead to a variety of presentations up to paraplegia below the level of the lesion along with urinary symptoms. A typical disk herniation will usually spare the exiting nerve root, but impinge upon the traversing nerve root of the level below. A rarer far lateral disk herniation, however, will impinge upon the exiting nerve root.

With age, the disk will degenerate. Autopsy specimens have noted disk degeneration starting as early as the second decade of life, and nearly all individuals have some degree of degeneration by the sixth decade. Osteophytes may then form around the disk space and cause stenosis of the spinal canal and neuroforamina.

Ninety-five percent of lumbar disk herniations occur at the L5/S1 and L4/L5 levels. Only 4% of lumbar disk herniations occur at the L3/L4 levels and are infrequent at the upper lumbar levels.

Clinical Findings

  1. Symptoms & Signs

The symptoms and signs of lumbar disk herniation are variable. A large central disk herniation can present with cauda equina syndrome. In these cases, patients may present with saddle anesthesia, urinary dysfunction, diminished rectal tone, and leg weakness. Generally, however, patients complain of symptoms of radiating leg pain with a variable component of back pain. Valsalva maneuvers (coughing, sneezing, etc) or movement will generally exacerbate the pain. Alternatively, rest will often improve the pain. The pain itself can be described as a constant burning, aching type pain with an intermittent sharp, shooting pain that radiates down into the legs. Straight leg raise and crossed straight leg raise may support the diagnosis of lumbar disk herniation. The straight leg raise is positive if raising the straightened leg to an angle of 30 degrees causes sciatica in the ipsilateral leg. This test is 80% sensitive but only 40% specific. The crossed straight leg raise is positive if raising the leg to 30 degrees causes sciatica in the contralateral leg, and though this test is only 25% sensitive, it is 90% specific. It should be noted that patients with a high lumbar disk herniation or a far lateral disk herniation may not have these signs. Examination of the paravertebral musculature may reveal tenderness and/or muscle spasm.

Motor findings may be helpful in predicting the involved lumbar level. Compression of the L4 nerve root (L3/4 herniation) may cause weakness of knee extension (quadriceps). Compression of the L5 nerve root (L4/5 herniation) may precipitate weakness of the extensor hallicus longus and ankle dorsiflexion (tibialis anterior). Finally, compression of the S1 nerve root (L5/S1 herniation) may lead to weakness of ankle plantarflexion (gastrocnemius). Reflexes may also be diminished. The ankle jerk (Achilles) reflex is diminished with S1 nerve root compression and the patellar reflex is diminished with L4 nerve root compression.

Sensory examination is often variable and the least helpful in predicting the involved lumbar level. L4 nerve root compression can be associated with anterior thigh to medial ankle sensory findings. L5 nerve root compression can present with findings along the dorsum of the foot and the first web space. Finally, S1 nerve root compression may present with sensory findings along the lateral and plantar regions of the foot.

  1. Imaging Studies

If symptoms are limited to pain and the patient does not have risk factors for other diseases, it is reasonable to delay imaging workup for 4 weeks since improvement in pain over time is not uncommon. Persistent symptoms, however, are an indication for imaging. Plain radiographs have limited utility in the diagnosis of disk herniation. However, they are useful in evaluating trauma, infection, or neoplastic process. Myelography may identify extradural filling defects and can be particularly helpful when combined with CT scanning. Indeed, CT myelography remains useful in the setting when an MRI scan is not possible.

MRI has become the gold standard for the diagnosis of herniated disks (Figure 36–26). MRI is noninvasive and does not involve radiation exposure. MRI provides detailed images of the disk spaces, surrounding soft tissue, and thecal sac. MRI can help exclude tumors, cysts, and postoperative scarring as etiologies of the patient’s symptoms. It is important to correlate the patient’s symptoms and the imaging findings precisely since MRI imaging can generate a significant number of false positives. For example, nearly 20% of normal individuals under the age of 40 and over 50% of individuals over the age of 40 were noted to have lumbosacral imaging abnormalities.

Figure 36–26. Sagittal T2-weighted MRI of lumbar spine showing multilevel disk bulges, ligamentum flavum hypertrophy, and retrolisthesis at L2-L3.

  1. Special Examinations

EMG can be useful diagnosing radiculopathy, but its utility is rather limited. EMG classically has findings of fibrillations at rest in the muscles supplied by a single nerve root and denervation of corresponding paraspinal muscles. Unfortunately, fibrillations require at least 3-4 weeks from the onset of the radiculopathy to be evident on EMG examination. Nerve conductions studies (NCS) are of minimal utility in diagnosing radiculopathy.

Differential Diagnosis

It is important to obtain a complete history and physical, as the differential diagnosis for patients with back pain and radicular symptoms is broad. A history of trauma can point to fractures, especially in the setting of osteoporosis and/or steroid use. Tumors which often metastasize to the spine include prostate, breast, kidney, thyroid, and lung cancer. Patients with metastatic disease often have nocturnal pain and pain that persists even with rest and a supine position. Inflammatory disorders, infections, bony abnormalities (spondylolithesis), peripheral neuropathies, degenerative spinal cord lesions, peripheral vascular occlusive disease, and peripheral nerve lesions should all be considered in the differential diagnosis.

Treatment

  1. Medical Measures

The natural history of the radicular pain associated with lumbar disk disease is that of improvement over time. Therefore, conservative measures are recommended for patients who present with a new radiculopathy without neurologic impairment. Conservative measures are directed toward initially limiting physical activity, including a brief period of bed rest followed by a gradual exercise program. It is also important for patients to modify their types of movement, for example to limit heavy lifting, twisting, or bending. Physical therapy can be useful after the acute period for instruction in abdominal and back musculature strengthening exercises.

The core of medical treatment is nonsteroidal anti-inflammatory drugs (NSAIDs). Oral steroids (ie, solumedrol dose pack) also may be useful in the acute setting, and epidural steroid injections and narcotics may be helpful in alleviating pain.

  1. Surgical Treatment

Patients who present with acute neurologic motor deterioration warrant immediate surgical attention. Surgery is also indicated in patients who fail the conservative measures outlined above and continue to suffer from debilitating pain.

The microdiskectomy is the “gold standard” surgical intervention for patients with a herniated lumbar disk. A microdiskectomy involves a laminotomy to gain access to the disk space. The nerve root and the thecal sac are protected while the disk fragment is identified and removed. A fusion can sometimes be recommended in the setting of recurrent disk herniations at the same level or pain associated with joint instability.

Prognosis

Overall, patients who have symptoms of radicular pain without neurologic deterioration have an excellent prognosis with improvement in symptoms over time. However, if loss of motor strength has already occurred, it is less likely to return even after surgical correction.

Boselie TF, Willems PC, van Mameren H, de Bie R, Benzel EC, van Santbrink H: Arthroplasty versus fusion in single-level cervical degenerative disc disease. Cochrane Database Syst Rev2012;9:CD009173.

Jacobs W, Van der Gaag NA, Tuschel A, et al: Total disc replacement for chronic back pain in the presence of disc degeneration. Cochrane Database Syst Rev 2012;9:CD008326.

Kovacs FM, Urrutia G, Alarcon JD: Surgery versus conservative treatment for symptomatic lumbar spinal stenosis: a systematic review of randomized controlled trials. Spine 2011;36:E1335-E1351.

Matz PG, Anderson PA, Holly LT, et al: The natural history of cervical spondylotic myelopathy. J Neurosurg Spine 2009;11: 104-111.

CSF DIVERSION FOR HYDROCEPHALUS

Hugh J. L. Garton, Jason Sack

OVERVIEW, EPIDEMIOLOGY, & PATHOPHYSIOLOGY

Hydrocephalus is a common diagnosis in both adult and pediatric patients. Most often this disease is chronically treated with an implanted catheter system to divert CSF from the brain to an alternative absorptive space such as the pleural or peritoneal space. Children and adults with CSF shunts can require surgical treatment for other associated conditions. The presence of the shunt may thus complicate the surgical management of various intra-abdominal processes. For example, a child with the distal portion of his shunt within the peritoneal may develop appendicitis. What management steps should be taken with regard to the management of the potentially contaminated CSF shunt catheter? In addition, CSF shunt failure is common, occurring in up to 30%-35% of individual within 1 year of initial shunt placement. About 1% of all shunt failures are fatal. Familiarity with the diagnosis and treatment of shunt failure is therefore desirable for providers caring for patients in whom a CSF shunt is present.

In pediatric patients, hydrocephalus is commonly seen in patients with a history of premature birth, intraventricular hemorrhate, after meningitis, in patients with myelomeningocele or other congenital cranial malformations and in patients with brain tumors. In adults, patients suffering from subarachnoid hemorrhage, brain tumor, or head injury may develop hydrocephalus. In older adults, “normal pressure” hydrocephalus is a potentially treatable cause of dementia. CSF shunt management is also used in the management of idiopathic intracranial hypertension (IIH) (also know as pseudotumor cerebri).

Among children, the prevalence of hydrocephalus is estimated at about 1-2 per 1000 children. Among adult patients, the incidence of IIH, and the shunt treatment for it appear to be increasing in accordance with increasing obesity rates.

Pathophysiologically, hydrocephalus results from an interruption in CSF circulation. The choroid plexus of the brain’s ventricular system generates about 80% of the total adult CSF production of about 20 cc/h, by an active ion pump dependent process. The remainder is thought to be generated by more general metabolic processes of the brain and arachnoid. Importantly, the production of CSF is independent of the intracranial pressure over a wide range of physiologic values. Thus, the increase in intracranial pressure that typically results from the increasing CSF volumes within the nervous system does not act to check the further production of spinal fluid. Once produced, spinal fluid moves in an oscillatory fashion out of the ventricular system and into the subarachnoid space around the brain. From there, it is reabsorbed by passive, pressure dependent mechanisms into the cerebral venous sinuses and possibly into lymphatic systems adjacent to the dura. The vast majority of hydrocephalus results from interruption of the egress of CSF. Traditionally, hydrocephalus has been classified as either obstructive, if the blockage to CSF outflow prevents egress from the ventricular system, or communicating if reabsorption is interrupted beyond the ventricular system, at the dural venous and/or lymphatic absorption sites. The determination of the site of obstruction is usually radiographic and may require invasive studies. To the non-neurosurgeon, this distinction is mostly important in determining the safety of lumbar puncture. If a patient has obstructive hydrocephalus, a lumbar puncture (LP) may be unsafe because of the potential for differential pressures between the cranial and spinal spaces after the LP. Radiographically, hydrocephalus is sometimes assumed to be communicating when all four ventricles are dilated, as opposed to only ventricles I-III. However, blockage at the outflow of the fourth ventricle to the subarachnoid space could produce the same CT or MRI picture despite obstructive physiology. Consultation with a neurosurgeon or neurologist may be helpful if the situation is unclear.

CLINICAL PRESENTATION/ASSESSMENT

The relative difference between the volume of CSF produced and reabsorbed, along with the relative compressibility of the brain determine the severity and type of clinical symptoms and signs at presentation. Slower buildup of CSF or a brain more compliant to compression, such as might occur at the extremes of age, produce a more protracted course of symptoms compared to rapid CSF accumulation in a poorly compressible brain. Whether presenting initially or after treatment failure, the symptoms of hydrocephalus can be grouped in to three broad categories: those related to acute increased intracranial pressure, usually seen with a more rapid accumulation of CSF or a poorly compliant brain, those related to more chronic deformation of the nervous system by more slowly accumulating CSF, and those symptoms that are specific to treatment complications, including CSF shunt infection.

Acute, rapidly progressive hydrocephalus presents with headaches, nausea, and vomiting and as it progresses, a deterioration in level of consciousness. The headaches associated with hydrocephalus and/or CSF shunt failure may have morning predominance, or worsen with a valsalva maneuver. Patient may be lethargic, or difficult to arouse or awaken with ordinary stimuli. Some patients may present with a loss of upgaze or Parinaud’s syndrome, others with cranial nerve VI palsy from increased intracranial pressure. In young children with an open fontanelle, this area may be tense and raised. Hydrocephalus of a more slowly progressive nature may present with more subtle signs of cognitive impairment. In the elderly, so called Normal Pressure Hydrocephalus presents with the triad of dementia, gait disturbance and incontinence. Other signs include papilledema, and in young children, an inappropriately expanding head size. If the patient has already undergone treatment for hydrocephalus with a CSF shunt, then additional symptoms and signs to consider include those of device infection, including a stiff neck, fever, and redness around the device. If the device has broken or malfunctioned then CSF may accumulation around the shunt. If the distal cavity into which a shunt is placed fails to absorb spinal fluid, there may be symptoms related to this such as abdominal pain with a large intraperitoneal fluid collection (so-called pseudocyst). Some CSF shunt devices possess diagnostic chambers that can be manually compressed and observed for response. However, the utility of such tests is questionable and best left to a neurosurgeon.

Patients presenting with clinical features described above, especially with a history of a CSF shunt in place for prior treatment of hydrocephalus must be promptly evaluated with imaging studies, usually a CT or MRI scan of the brain. In the setting of a possible shunt failure, it is critical to have previous images available for comparison. Patients with working shunts may have ventricles that are smaller than normal; “normal ventricular size” or “no evidence of shunt failure” in a radiology report has been demonstrated to correlate poorly with final diagnosis. Additional studies that can be useful include plain radiographs of the shunt and imaging of the cavity into which the CSF shunt is draining, such as by ultrasound of the abdomen. Other invasive diagnostic studies may be necessary, usually as directed by a neurosurgical consultant. These can include a CSF shunt tap, in which a portion of the shunt that sits under the scalp is accessed percutaneously similar to a vascular access port. It is helpful to know that a high percentage of shunts failures occur within the first 2 years after shunt failure, with a reduction in failure rates following this. In children, young age at presentation also appears to be a risk factor for repeated shunt failure. Similarly, most shunt infections occur within 1 year from shunt placement. However, despite these epidemiologic data, a perfectly predictive clinical decision algorithm remains elusive and a low threshold for obtaining imaging studies and appropriate expert consultations is warranted.

MANAGEMENT

Progressive hydrocephalus must be dealt with in a timely manner to minimize neurologic deterioration. As noted above, the rate of pathologic CSF accumulation may vary and produce different clinical symptoms. However, because of the potential for acute and fatal deterioration from hydrocephalus, the initial presumption should be that rapid intervention is needed, subject to reconsideration after all the data are available.

Medical management of hydrocephalus may be appropriate in the management of IIH (pseudotumor cerebri), and in some patients after subarachnoid hemorrhage. It is also used in the initial management of hydrocephalus in neonates after an intraventricular hemorrhage. The diuretics acetazolamide and furosemide are used in this context. Both decrease CSF production by inhibiting carbonic anhydrase, albeit by slightly different mechanisms—thus producing an additive effect when used in combination. However, randomized control trials of diuretic therapy in newborns demonstrated no reduction in the need for subsequent shunt placement for patients on diuretic regimens. If diuretics are utilized, patients must be closely monitored for electrolyte imbalances and acetazolamide toxicity (acute gastritis, parasthesias, drowsiness.). It should be noted that with the exception of IIH, protracted use of diuretics to manage hydrocephalus is rarely successful and in most cases, surgical therapy is indicated without a trial of diuretics.

Intraventricular administration of fibrinolytic agents (eg, streptokinase) has also been investigated as a treatment for newborns with posthemorrhagic hydrocephalus. However, a systematic review of randomized trials failed to show any benefit, in terms of reducing shunt requirement or death. Furthermore, secondary intraventricular hemorrhage is a potential complication of this treatment.

Temporary drainage includes serial LPs, ventricular taps, placement of an external ventriculostomy catheter, or implanted reservoir in communication with the ventricular system for periodic aspiration. These techniques are used when natural history suggests possible resolution of the hydrocephalus, as may occur in IVH in the newborn and in SAH. In addition, temporary diversion is appropriate when the patient’s condition precludes definitive treatment.

Endoscopic third ventriculostomy (ETV) is an alternative to CSF shunt placement for the treatment of some forms of hydrocephalus. During the endoscopic procedure a perforation is made in the floor of the third ventricle communicating it to the subarachnoid space. ETV is currently used as the initial treatment of choice in cases of obstructive hydrocephalus and is curative in 80% of properly selected patients, avoiding the need for shunt placement. ETV patients may suffer recurrent hydrocephalus if the fenestration closes. The vast majority of such failures have occurred within the first year after the procedure, and failures beyond 5 years are exceedingly rare.

CSF shunt treatment is indicated in most cases of progressive hydrocephalus, when conservative management and ETV are not indicated, or have previously failed. This procedure involves placement of a mechanical shunt as a means to divert excess CSF from the ventricles into other body cavities. CSF shunts have a ventricular catheter, a valve that regulates unidirectional flow, and distal tubing to distribute the CSF to its absorption site. Many shunts also have reservoirs that can be percutaneously tapped for diagnostic purposes, or as a temporizing measure in shunts in which the distal tubing or valve becomes blocked. Most often, the proximal catheter of the shunt system is placed into one of the lateral ventricles, either in the frontal horn via a frontal approach or via the atrium of the lateral ventricle via a parietal approach. Shunt systems are used to treat a variety of other conditions such as intracranial cysts, or chronic subdural hematomas/hygromas. The proximal catheter placement in such cases is dictated by the location of the pathology to be treated. Shunt valve mechanics regulate the amount of drainage. Most are differential pressure systems that respond to increasing fluid pressure by allowing more CSF through the valve. Various shunts are deigned to open in different pressure ranges (eg, low: 4-7 cm H2O; medium: 8-12 cm H2O; high: 13-15 cm H2O), and some valves can be externally adjusted to different performance levels using a magnet. These adjustable (“programmable”) valves may be inadvertently reprogrammed by other external magnets and while all systems approved for sale in the United States as of this writing are MRI compatible to 1.5 Telsa, the patient must have the valve setting checked and reset as necessary following an MRI. Alternative systems are designed to produce a constant flow through the valve over a range of physiological intracranial pressures. Valve systems may incorporate devices that prevent excessive drainage from the siphoning effect of a long run of distal tubing running inferiorly when patients assume an upright posture. The distal end of the system is comprised of a catheter that terminates in a body cavity, wherein the fluid can be adequately absorbed. A ventriculoperitoneal (VP) shunt, in which the catheter terminates into the abdominal cavity and fluid is absorbed by the peritoneum, is the most commonly used shunt at present. Other common distal sites include the pleural space (V-Pleural shunt) and the cardiac atria (VA shunt). These two sites are chosen when extensive scarring or adhesions, recent abdominal infection, peritonitis, or morbid obesity preclude peritoneal catheter placement. Other less common options for distal placement of a ventricular shunt include gall bladder, and ureter/bladder. Immediate complications from shunt placement are fortunately rare but include misplacement of the ventricular catheter with inadequate shunt function. Much more rarely a patient may suffer intracranial hemorrhage or perforated abdominal viscus.

Complications that are not encountered intraoperatively, but may nonetheless present in the early postoperative period include: inguinal hernia and/or hydrocele; ascites; pseudocyst formation; septicemia, pulmonary thromboembolism, and cardiac tamponade with VA placement; subcutaneous CSF collections and fistulas; hemorrhage; shunt obstruction/occlusion; and infection (shunt and/or incisional).

CSF SHUNT FAILURE

While the surgical procedure of CSF shunt placement is generally an uncomplicated process, the long-term management of shunted hydrocephalus is more problematic. Multiple clinical studies in both adult and pediatric patients attest to the high rate of device failure, from tubing obstruction, catheter fracture, infection, excessive drainage, and compartmentalization of the ventricular system (with a catheter draining part of the ventricular system, but with another part expanding because of noncommunication). In young children undergoing first shunt placement, roughly 1/3 of shunts will require reoperation in the first year following surgery. Shunt infection rates in these children are as high as 10%-12%. Adult patients fair somewhat better but 20% will still require reoperation within the first year after shunt failure. Analysis of the “shunt survival” curves shows that failure rates for devices drops considerably after the first year or two from surgery, but the threat of failure remains present to some degree as long as the device remains in place.

In evaluating a patient for potential CSF shunt failure, the epidemiological data above provide a baseline estimate of the probability of the diagnosis. Other risk factors for shunt failure include young age at shunt placement and recent previous shunt surgery. A history and physical exam along with images compared to previous findings, as described above, should allow for a reasonably accurate diagnosis. However, several pitfalls deserve mention. First, as has been noted, not all patients with CSF shunt failure will show significant expansion of the ventricular system. Some patients, particularly those with a long history of shunted hydrocephalus may present with the so-called “slit-ventricle” syndrome. The radiographic finding of “slit-ventricles” is relatively common for patients with a functioning shunt. The clinical “slit-ventricle” syndrome presents with episodic severe headaches that appear to be due to intermittent occlusion of the shunt in a small ventricular system that does not dilate despite an increase in CSF pressures because change in ventricular size is too small to be noted on standard radiographs. The diagnosis is often confirmed by direct intracranial pressure measurements obtained from a separately placed ICP monitor.

A second group that deserves special attention is children and adults with myelomeningocele. Given a frequent need for both urological and plastic surgical care, these patients are frequently encountered on surgical services. About 70% of patients with this spinal dysraphism will require treatment for hydrocephalus, most with a CSF shunt. Myelomeningocele is associated with a number of abnormalities of the brain stem and foramen magnum. These abnormalities allow CSF shunt failure to produce a much wider array of clinical symptoms than would otherwise be the case. As intracranial CSF and pressure build because of CSF shunt failure, downward pressure on the brainstem can produce lower cranial nerve palsies, including swallowing and breathing irregularities. The death rate from shunt failure may be higher in this population, and more precipitous, related to the susceptibility of the hindbrain to herniation. In additional, CSF shunt failure may precipitate or exacerbate a preexisting syrinx in these patients, producing symptoms related to spinal cord dysfunction.

Shunt contamination may occur during unrelated intra-abdominal procedures, in the case of a VP shunt, or from repeated bacterema in ventriculoatrial shunts. Infections of this sort can obviously occur at any time, and are not temporally restricted to the year or so following the most recent shunt surgery. Clinically, the presentation of infection may lack an overt inflammatory response with fever and redness around the shunt. Rather, the presentation may be more insidious, with failure of the CSF to absorb and development of a pseudocyst. The presence of a large intra-abdominal fluid collection in a patient with a VP shunt should raise suspicion of shunt infection, although sterile fluid collections are not uncommon.

Abdominal surgical procedures in patients with VP shunts raise the issue of the best management of an intraperitoneal distal shunt catheter during and after the procedure. While no validated guideline exists, the literature and the author’s experience suggest that the infection risk is low during clean and clean contaminated cases. If an enterotomy or open bladder procedure is to be performed, it is reasonable to relocate the shunt away from this area once the tubing is identified intraperitoneally, and/or protect it with gause sponges during the procedure. Other than routine preoperative antibiotics, expectant management for shunt infection and failure can be practiced and the shunt does not require externalization. In situations where there is frank contamination of the shunt catheter in the abdomen, externalization of the shunt appears prudent. Laparoscopic procedures have been demonstrated to transiently increase intracranial pressure, but there is no body of literature to suggest that this produces identifiable complications, presumably due to its transient nature. Constipation and ileus, conversely have been suggested to be a source of at least transient shunt dysfunction due to increased abdominal pressure.

Shunt externalization, when necessary, involves palpation of the shunt proximal to its entry into the peritoneum, followed by a sterile prep and, if the patient is awake, local anesthesia. If the shunt has been in place for several years, it may be quite adherent to the surrounding tissues and the procedure is then best performed in the operating room with anesthesia. For recently placed shunts, this is less of an issue and the procedure can be performed at the bedside, assuming a cooperative patient. A cut down is made of over the shunt tract, avoiding laceration of the underlying tubing. The catheter is often encased in a thick sheath of scar that must be teased open to gain access to the catheter. If a specimen of fluid is required from a peritoneal fluid collection, the tubing is withdrawn slightly, then cut and aspirated distally. In a large pseudocyst a liter or more of fluid may be withdrawn. The distal catheter is then withdrawn from the patient and then discarded. The remaining tubing should be observed for CSF drainage and connected to a sterile, enclosed CSF drainage system. Neurosurgical consultation is advisable before embarking on the procedure. Patients with externalized shunts are draining isotonic CSF and may suffer hyponatremia if appropriate fluid and electrolyte replacement are not performed, particularly in young children.

ONGOING CARE

Patients with VP or ventriculopleural shunts do not require prophylactic antibiotics for surgical or dental procedures in which the shunt is outside the operative field. Prophylactic antiobiotics may be helpful when the shunt is in the operative field, as described above, or during a planned CSF shunt revision. For patients with ventriculo-atrial CSF shunts in place, an argument can be made for antibiotic prophylaxis before procedures likely to result in bacteremia including dental procedures. Dental practice guidelines suggest this also. American Heart Association guidelines for endocarditis prophylaxis recommend Amoxicillin 2 g 30-60 minutes before the procedure for adults and 50 mg/kg for pediatric patients or Clindamycin 600 mg adult and 20 mg/kg pediatric doses for penicillin allergic patients.

Shunt independence has been reported for children with a prior history of shunt placement. In this situation, a child previously depended on a CSF shunt becomes once again able to drain CSF independently. Clinically, this issue is often raised when a CSF shunt has gone without a revision for many years or even decades, or a CSF shunt is found disconnected in the absence of clinical symptoms. One series reported that 3% of children with hydrocephalus became shunt independent later in life. However, in the setting of a shunt found to be fractured on x-ray without ventricular enlargement or clinical symptoms of shunt failure, it is clinical experience that CSF can often drain between the fractured catheter segments through the tube of scar tissue that forms around the catheter initially after shunt placement. Such a scarred tract can subsequently close over, even several years after the catheter fracture and separation of the fractured tubing ends. Therefore, caution should be exercised in the presumption of shunt independence unless it has been verified by invasive testing. Patients incidentally found to have broken shunt catheters may reasonably be referred to a neurosurgeon for evaluation.

Shunted hydrocephalus during pregnancy frequently raises concerns over management of headaches during pregnancy, mechanism of delivery, and over the potential impact of hydrocephalus on the pregnancy outcome. Pregnancy, particularly in the third trimester, increases intra-abdominal pressures. For patients with VP shunts, this can lead to a relative decrease in CSF shunt function. In cases series of pregnant patients with hydrocephalus, headaches are not uncommonly reported, particularly during the third trimester. These may herald shunt failure, the evaluation of which may be complicated by a desire to avoid radiation administration during pregnancy. MRI may be a useful option in these circumstances. In the absence of radiographic manifestations of shunt failure, or additional clinical symptoms, safe observation of headaches has been successfully practiced, but treatment decisions must be individualized and other diagnoses, particularly eclampsia and preeclamsia must be considered. There is little evidence to suggest that the presence of a CSF shunt is a contraindication to labor and vaginal delivery. It has been argued that patients who are suspected of being symptomatic from increased intracranial pressure may benefit from avoiding protracted labor, but no comparative studies exist to provide a clear answer. Hydrocephalus, and the presence of a CSF shunt does not, per se, appear to impact pregnancy outcome. However, recalling that hydrocephalus is coincident with other diagnoses, such as epilepsy and myelomeningocele that do have a significant impact on the potential for birth defects, the importance of prenatal care can be emphasized.

Whitelaw A and Odd DE: Intraventricular streptokinase after intraventricular hemorrhage in newborn infants. Cochrane Database Syst Rev 2007;(4):CD000498.

Li G, Dutta S.: Perioperative management of ventriculoperitoneal shunts during abdominal surgery. Surg Neurol 2008;70(5): 492-495.

Wilson W, et al: Prevention of infective endocarditis: guidelines from the American Heart Association: a guideline from the American Heart Association Rheumatic Fever, Endocarditis, and Kawasaki Disease Committee, Council on Cardiovascular Disease in the Young, and the Council on Clinical Cardiology, Council on Cardiovascular Surgery and Anesthesia, and the Quality of Care and Outcomes Research Interdisciplinary Working Group. Circulation 2007;116(15):1736-1754.

Wu Y, et al: Ventriculoperitoneal shunt complications in California: 1990 to 2000. Neurosurgery 2007;61(3):557-562; discussion 562-563.

CENTRAL NERVOUS SYSTEM INFECTIONS

Khoi D. Than Anthony C. Wang Jean-Christophe A. Leveque Stephen E. Sullivan

BRAIN ABSCESS

Brain abscesses are an uncommon entity, with approximately 2000 cases reported in the United States each year. There is a higher incidence in developing countries, and men are affected slightly more often than women. Classically, these abscesses arise locally from otorhinolaryngeal infections or hematogenously from distant infections, though opportunistic infections have become an important consideration upon initial presentation as well. The pathogenic organisms most commonly implicated are of the Streptococcus family; Klebsiella, Staphylococcus aureus, and anaerobes are also frequent. In immunocompromised patients, it is important to include Toxoplasma, Listeria, and Nocardia as possible etiologic agents, as well as fungal pathogens.

A patient with a brain abscess can present with nonspecific symptoms. Headache, nausea, vomiting, and altered mental status can occur due to increased intracranial pressure (ICP), while unilateral headache, seizures, and many focal neurological deficits occur due to the presence of a mass lesion. Fever and nuchal rigidity are also seen in many cases. Additional findings in the newborn patient may include cranial enlargement, meningeal signs, irritability, and failure to thrive.

Risk factors for brain abscess include sinus, ear, or dental infections. These sources usually lead to formation of frontal or temporal lobe abscesses through direct spread. Hematogenous spread from intra-abdominal, pelvic, pulmonary, or cardiac seeding occurs most commonly via the middle cerebral artery, leading to microembolic infarcts at the gray-white junction. Risk factors for these types of abscesses include infectious lung processes and congenital cyanotic heart disease. In these conditions, the lungs have a decreased filtering capability, and the associated relative hypoxia promotes abscess formation. Head trauma—blunt, penetrating, or surgical—can introduce a nidus for infection with delayed abscess formation. Parasitic infections such as cysticercosis should be considered more likely in recent foreign travelers.

The differential diagnosis for brain abscess includes subdural empyema, septic emboli, dural sinus thrombosis, mycotic aneurysm, meningitis, focal necrotizing encephalitis (herpes simplex virus), and tumor, as all of these conditions can present with headaches and altered mental status. In the initial evaluation of brain abscess, blood work that should be drawn includes a white blood cell count, cultures, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP); however, normal test results do not rule out the diagnosis. The key to diagnosing brain abscess is correlating the clinical scenario with an imaging study, such as contrast-enhanced computed tomography (CT) or MRI. The classic finding on CT or MRI is a circular lesion with a strongly contrast-enhancing surround rim. CT images are typically the first obtained on admission, although MRI is the imaging modality of choice as it can provide greater anatomic detail. MRI evaluation for brain abscess should always include diffusion-weighted images (DWI), which can differentiate between ring-enhancing lesions of infectious and neoplastic origin, as abscesses are typically hyperintense on DWI while neoplastic lesions are hypointense.

One general warning is to avoid LP, as cerebrospinal fluid (CSF) results are often nondiagnostic and this procedure is associated with a worsened outcome in patients with brain abscesses. Less than one-quarter of patients have positive CSF cultures, and, with a large enough abscess, there is a real risk for transtentorial or brainstem herniation. CSF sampling should be considered only if parasitic pathogens are suspected. A definitive diagnosis is made by biopsy sampling of the abscess through surgical means.

The treatment of brain abscesses involves both surgical and medical therapy. Treatment should also be aimed at correcting the primary source of infection (ie, draining a pulmonary empyema or repairing a correctable heart defect). Initial surgical treatment usually consists of needle aspiration of the abscess; the abscess can be monitored serially with imaging, and the procedure repeated as necessary. A total excision via a craniotomy can be performed if the abscess is in its chronic, encapsulated form and is located in a surgically amenable region of the brain. It is advisable to perform surgery before starting antibiotics in order to confirm the diagnosis as well as to identify the organisms and their antibiotic sensitivities. Antibiotic therapy typically consists of 6-8 weeks intravenous treatment followed by 4-8 weeks oral treatment. Patients should receive routine follow-up imaging and should also be started on an antiepileptic medication. Glucocorticoids should be considered to counteract symptomatic intracranial hypertension, although their role is less important than in the treatment of brain tumors.

In certain situations, medical therapy can suffice without the need for surgery. These situations include an abscess in its early stages (ie, symptoms for less than 2 weeks), a small (< 2 cm) abscess, or a definite clinical improvement after 1 week of antibiotics only. Medical treatment without surgery should also be considered in poor surgical candidates, patients with multiple abscesses and/or concomitant meningitis, patients with abscesses in eloquent locations, or patients with hydrocephalus and ventricular shunts.

Patients with brain abscess have a reported mortality risk of 0%-30% depending on etiology and presentation. An overall 50% morbidity risk of permanent neurological deficits is conferred, which depends heavily upon the severity of presenting symptoms.

SUBDURAL EMPYEMA

A subdural empyema is a collection of pus that forms in the subdural space. It is less common than brain abscess, but like abscesses, it is more commonly found in men. Subdural empyema is an emergent condition because, unlike the brain parenchyma with abscesses, the subdural space does not pose much of a barrier to prevent the spread of infection. Additionally, antibiotics have poor penetration into the subdural space.

The most common cause of subdural empyema (70%) is paranasal sinusitis, especially in cases involving the frontal sinus. Chronic otitis media accounts for another 15% of cases. As such, the organisms typically cultured from a subdural empyema include Streptococcus (aerobic and anaerobic) and Staphylococcus. Symptoms present in the majority of patients with subdural empyema include fever, headache, nuchal rigidity, hemiparesis, and altered mental status. Other common symptoms include seizures and sinus tenderness.

Computed tomography or MRI imaging will typically diagnose a subdural empyema. Three-fourths of empyemas are located over the convexity, while 15% are parafalcine (ie, adjacent to the falx cerebri). Just as with brain abscesses, LP should be avoided due to the risk of herniation.

Almost all cases of subdural empyema will require surgical drainage, preferably emergently. The two surgical options are burr-hole drainage and craniotomy. Although burr-hole drainage is less invasive, it is also less effective; thus, craniotomy is the preferred surgical option. Antibiotics are used for a course of 4-6 weeks, and patients are put on therapeutic or prophylactic antiseizure medication. Medical treatment alone can be effective if the empyema is small, there is minimal neurologic involvement, and antibiotics have an early efficaciousness.

Subdural empyema carries a 15% mortality rate. Half of patients have residual neurological deficits at the time of hospital discharge. Factors known to be associated with poor prognosis include age more than 60 years, obtunded or comatose state at presentation, and empyema formation secondary to surgery or trauma.

OSTEOMYELITIS

Osteomyelitis can affect the skull or the vertebrae. Osteomyelitis of the skull usually results from contiguous spread from an infected sinus or from penetrating trauma (ie, postoperative). The infectious agents are typically S. aureus or epidermidis, and treatment consists of debridement surgery followed by 6-12 weeks of antibiotics (intravenously for the first 1-2 weeks). Surgical treatment is aimed at removing all infected bone. A cranioplasty or other hardware is not placed until several months later in order to minimize the risk of reseeding an infection.

Vertebral osteomyelitis (VO) represents 3% of all cases of osteomyelitis, and is more common than skull osteomyelitis because of the spine’s rich vascular supply. Both anterograde arterial seeding as well as retrograde venous plexus spread have been implicated in VO, with S. aureus as the most common organism. VO caused by Mycobacterium tuberculosis is known as Pott’s disease. Those at higher risk for developing VO include intravenous drug users, diabetics, sickle cell patients, patients on hemodialysis, and the elderly.

The most common presentation in patients with VO is that of back pain (> 90%), usually unaffected by activity. Other typical presenting symptoms include fever, weight loss, radicular pain, and myelopathy. The neurologic symptoms are usually a result of destruction of the vertebral body and subsequent retropulsion of bone into the spinal canal or neural foramen. The most commonly affected segment of the spine is the lumbar region followed by, in order, the thoracic, cervical, and sacral segments. Any source of infection can theoretically put one at risk for developing VO, although important sources include infections of the urinary tract, respiratory system, and mouth. VO also develops at sites of previous spine surgeries.

Definitive diagnosis of VO is made with positive cultures, either from biopsy of the tissue itself or via blood cultures in the setting of suggestive radiographic findings. MRI has demonstrated excellent diagnostic accuracy of more than 90% of cases and is the preferred diagnostic modality. When unable to obtain MRI, bone scintigraphy with single photon emission computed tomography has also demonstrated excellent sensitivity.

The treatment of VO is nonsurgical in the vast majority of cases, with disease resolution being accomplished via antibiotic therapy alone. The goal of therapy should be to minimize neurologic involvement, and to maintain structural stability of the spine. Surgical treatment is indicated to obtain a tissue diagnosis if closed needle biopsy is unfeasible. In patients with a worsening neurologic deficit, the onset of structural instability, or a failure of medical management, surgery is warranted for abscess drainage, alleviation of compression, and stabilization.

SPINAL EPIDURAL ABSCESS

Spinal epidural abscesses (SEA) are often associated with VO, with the majority of cases arising from S. aureus. Streptococcus species are the second most commonly implicated organism (Figure 36–27). SEA are located most often in the thoracic region (50%), followed by the lumbar (35%), and cervical (15%) regions. The vast majority of abscesses (80%) are located posterior to the spinal cord.

Figure 36–27. Sagittal magnetic resonance T1-weighted image with contrast of the lumbar spine demonstrating diskitis/osteomyelitis associated with a spinal epidural abscess.

The primary infection leading to SEA can be from hematogenous spread or direct extension. Hematogenous spread is more common, with skin infections being the usual originating source. Other mechanisms of hematogenous spread include nonsterile intravenous injections, bacterial endocarditis, urinary tract infections, respiratory infections, and oropharyngeal abscesses. SEA caused by direct extension can be from decubitus ulcers or penetrating trauma, including following spinal procedures.

Patients with SEA are typically middle aged. Risk factors for developing SEA include diabetes, intravenous drug use, chronic renal failure, and alcoholism. Patients often present with back pain, spine tenderness, fever, sweats, and rigors. When motor weakness ensues, there is a very rapid progression to paraplegia. Thus, the diagnosis and treatment of SEA is emergent.

The workup of SEA should include a complete blood count, ESR, CRP, and blood cultures. An LP is contraindicated given the potential to spread infection from the epidural to the intradural space. The imaging modality of choice is MRI, although CT and myelography may also be used to arrive at a diagnosis.

The treatment of SEA, as with most infections of the central nervous system, is surgery plus antibiotics. Surgery is used to drain pus, debride any granulation tissue, and provide stability (usually in cases where there is bony destruction secondary to VO). Antibiotics are given intravenously for 3-4 weeks and then orally for another 4 weeks. Nonsurgical management with antibiotics only is rare and reserved only for very poor surgical candidates, abscesses that are very extensive in length, or cases in which complete paralysis has been present for at least 3 days with irreversible neurologic injury.

The overall prognosis for patients with SEA is relatively poor, with a mortality rate of 20%. In patients that survive, restoration of baseline neurological function is rare.

MULTIPLE CHOICE QUESTIONS

  1. Intracranial pressure
  2. Is normally between 30 and 40 torr
  3. Is directly (linearly) related to increasing intracranial mass
  4. Cannot be measured directly
  5. Is normally maintained at a stable level by displacement of CSF
  6. Is often affected by changes in the size of the skull after trauma
  7. Strategies to reduce intracranial pressure can include all of the following except
  8. Drainage of CSF
  9. Hyperventilation
  10. Treatment with mannitol
  11. Trendelenburg position
  12. Sedation
  13. Spinal cord injury
  14. Rarely includes the use of systemic corticosteroid therapy
  15. Is accompanied by initial hyporeflexia
  16. Is termed complete if there is no motor function below the level of injury
  17. Is accompanied by priapism and increased anal sphincter tone
  18. Can cause a Brown–Séquard syndrome, with loss of motor function and loss of pain and temperature sensation below the level of the lesion, with preserved proprioception, vibration, and pressure sensation
  19. Peripheral nerve injury recovery
  20. Occurs with axonal regeneration after wallerian degeneration at a rate of 1 mm per day
  21. Is best treated by delayed (3 month) repair is the case of acute sharp injury
  22. Should be treated by segmental resection and nerve graft in cases of apparent stretch injury
  23. Occurs more quickly with systemic corticosteroid therapy
  24. Is likely to be functionally successful if a neuroma forms
  25. Pituitary tumors
  26. Are typically adenocarcinomas
  27. Should be treated by urgent operation in most patients
  28. Can cause symptoms related to the compression of the pituitary stalk causing increased prolactin levels
  29. Can cause Cushing syndrome by overproduction of growth hormone
  30. Cause visual symptoms due to hormonal release


If you find an error or have any questions, please email us at admin@doctorlib.org. Thank you!