Patients with Concussion and Transient Symptoms
Patients with an uncomplicated concussive injury who have already regained consciousness by the time they are seen in a hospital and have a normal neurologic examination pose few difficulties in management. They should not be discharged until the appropriate examinations (CT scans, skull films, if necessary) have been obtained and the results prove to be negative. Also, the patient should not be released until the capacity for consecutive memories has been regained and arrangements have been made for observation by the family of signs of possible, although unlikely, delayed complications (subdural and epidural hemorrhage, intracerebral bleeding, and edema). A program instituted by Mittenberg and colleagues (2001) has shown that reassurance and explanation of the concussive injury and anticipated aftereffects reduce the incidence of postconcussive symptoms at 6 months. Most such patients become mentally clear, have mild or no headache, and are found to have a normal neurologic examination. They do not require hospitalization or special testing, but in the current litigious climate of the United States, some form of brain imaging is nonetheless often performed as discussed in an earlier section.
Acetaminophen may be prescribed for headache. Any increase in headache, vomiting, or difficulty arousing the patient should prompt a return to the emergency department. A written instruction sheet with symptoms to be expected and clear advice about returning for examination is very helpful.
Patients with persistent complaints of headache, dizziness, and nervousness, are the most difficult to manage. The main approach is to counsel patients while the symptoms resolve, coupled with a reduction in mental and physical effort that is commensurate with the patient's level of endurance. A program must be planned in accordance with the basic problem. If work or school work precipitate headaches, for example, plans should be made to have them curtailed. Half-time work may be suitable for some individuals but not for others. Similarly, some physical activity is to be encouraged but exertion that causes headaches or mental confusion to occur or worsen should be reduced. At the same time, a bedbound or homebound state is discouraged and the patient may walk, use the internet, watch television, or read up to the level of causing fatigue. Each of these activities is then increased at a gradual rate. In all instances, reassurance that these symptoms improve over weeks or more should be offered in order not to allow the individual to internalize the notions of chronic dementia after head injury that pervade the popular press. Some hard-driving patients return to work, only to find headache, confusion, and fatigue recur in a disabling way and must start the cycle of reduced effort over again.
If there is mainly an anxious depression, antidepressant medications—such as in a fluoxetine, paroxetine, or a tricyclic—may be useful, but their effects are often disappointing. Simple analgesics, such as acetaminophen or nonsteroidal antiinflammatory drugs, should be prescribed for the headache. Litigation should be settled as soon as possible. To delay settlement usually works to the disadvantage of the patient. Long periods of observation, repetition of a multitude of tests, and waiting only reinforce the patient's worries and fears and reduce the motivation to return to work. Neuropsychologic tests may be useful in the group with persistent cognitive difficulty, but the results should be interpreted with caution, as depression and poor motivation will degrade performance.
Patients with Severe Head Injury
If the physician arrives at the scene of an accident and finds an unconscious patient, a rapid examination should be made before the patient is moved. First it must be determined whether the patient is breathing and has a clear airway and obtainable pulse and blood pressure, and whether there is hemorrhage from a scalp laceration or injured viscera. Severe head injuries that arrest respiration are soon followed by cessation of cardiac function. Injuries of this magnitude are often fatal; if resuscitative measures do not restore and sustain cardiopulmonary function within 4 to 5 min, the brain is usually irreparably damaged. Bleeding from the scalp can usually be controlled by a pressure bandage unless an artery is divided; then a suture becomes necessary. Resuscitative measures (artificial respiration and cardiac compression) should be continued until they are taken over by ambulance personnel. Oxygen should then be administered.
The likelihood of a cervical fracture–dislocation, which may be associated with any severe head injury, is the reason for taking precautions in immobilizing the neck and in moving the patient. In the awake patient, neck pain calls attention to this complication. It should be recalled that even in the absence of a spinal fracture, the spinal cord may be threatened by the instability resulting from ligamentous injuries (posing the risk of subluxation). In the study of 292 patients with traumatic cervical injuries by Demetriades and colleagues, 31 (11 percent) showed subluxations without fracture and 11 (4 percent) had cord injuries with neither fracture nor subluxation. The combined use of standard cervical spine films and cervical CT scanning detected all cervical injuries. After severe head or neck injury, even without direct impact to the neck, it is advisable to obtain standard anteroposterior, lateral, and oblique neck films, with additional gentle flexion (20 degrees) and extension (30 degrees) views of the neck and a neck CT scan. If these are normal and there is little or no neck pain, the cervical collar is no longer required. If after these studies, or if they cannot be obtained, or if there is significant persistent pain or other neurologic findings induced by head movement, a cervical MRI is advisable. If there are signs of a myelopathy such as flaccid legs or incontinence, urgent MRI is advisable.
In the hospital, the first step is to clear the airway and ensure adequate ventilation by endotracheal intubation if necessary. A search for other injuries must be made, particularly of the abdomen, chest, spine, and long bones. Chestnut et al, in analyzing the data from the Traumatic Coma Data Bank, found that sustained early hypotension (systolic blood pressure <90 mm Hg) was associated with a doubling of mortality. If shock was present on admission to the emergency ward, the mortality was 65 percent. Although the hypotension that follows most injuries is a vasodepressor response and usually comes under control within approximately an hour without pressor drugs, a large, unimpeded intravenous line should be inserted. Persistent hypotension because of head injury alone is an uncommon occurrence and should always raise the suspicion of thoracic or abdominal internal bleeding, extensive fractures, or trauma to the cervical cord, or diabetes insipidus. Initially, the infused fluid should be normal saline, avoiding the administration of excessive "free water" because of its adverse effect on brain edema. Oxygen should continue to be administered until it can be shown that the arterial oxygen saturation is normal.
A rapid neurologic survey can then be made, with attention to the depth of coma, size of the pupils and their reaction to light, ocular movements, corneal reflexes, facial movements during grimace, swallowing, vocalization, gag reflexes, muscle tone and movements of the limbs, predominant postures, reactions to pinch, and reflexes. Bogginess of the temporal or postauricular area (Battle sign), bleeding from the nose or ear, and extensive conjunctival edema and hemorrhage are useful signs of an underlying basal skull fracture. However, it should be remembered that rupture of an eardrum or a blow to the nose may also cause bleeding from these parts. Fracture of the orbital bones may displace the eye, with resulting strabismus; fracture of the jaw results in malocclusion and discomfort on attempting to open the mouth. If urine is retained and the bladder is distended, a catheter should be inserted and kept there. Temperature, pulse, respiration, blood pressure, arterial oxygen saturation, and state of consciousness should be checked and charted every hour. The Glasgow Coma Scale, mentioned earlier, has provided a practical means by which the state of impaired consciousness can be evaluated at frequent intervals (see Table 35-1), but it should not be considered a substitute for a more complete neurologic examination.
CT and MRI scanning of the cranium have assumed central importance at this juncture. A sizable epidural, subdural, or intracerebral blood clot is an indication for immediate surgery. The presence of contusions, brain edema, and displacement of central structures calls for measures to monitor progression of these lesions and to control intracranial pressure. These measures are best carried out in a critical care unit.
Management of Raised Intracranial Pressure
There has been a presumption, not unreasonable, that high levels of intracranial pressure are deleterious after head injury, much as it is in other processes that involve an intracranial mass. At issue has been the precise pressure at which damage occurs, whether lowering ICP improves outcome, which treatments are best, and the role of monitoring in guiding treatment. Certainly there are many biologic processes in neurons and astrocytes that greatly influence outcome after traumatic brain injury, many set in motion at the time of impact and not referable to raised ICP. At times, these overwhelm the changes induced by intracranial pressure but they are not particularly remediable, producing an emphasis on reducing intracranial pressure as a means of preventing secondary brain damage. An approach to intracranial pressure treatment is given here and also addressed in Chap. 17 on coma and Chap. 31 on brain tumors.
In cases of moderate and severe head injury it has been the practice on most ICU services to insert one of several available devices that continuously record intracranial pressure (ICP). The rationale is ostensibly to gain control over a remediable cause of secondary brain damage, particularly if the patient's neurologic examination is reduced to a few sentinel signs such as pupillary enlargement or because sedating medications have obscured the examination. The ventricular catheter has been considered a "gold standard" of pressure measurements as it is directly coupled to the CSF compartment, which should best reflect the summated pressures within the cranium. It has the additional advantage of affording therapeutic drainage of CSF in order to reduce ICP. In comatose patients, monitoring of ICP could avoid excessive fluid administration, refine the amount of osmotic agent and hypertonic saline used to reduce pressure, and establishes the ideal level of hyperventilation. In these respects, monitoring can be helpful by guiding treatment and avoiding detrimental effects on ICP of treatments for head trauma.
However, there are few critical data to support the routine use of ICP monitoring. Certainly in the patient who is only drowsy or shows only minimal mass effect on CT, it is usually not necessary. Guidelines given by the American Association of Neurological Surgeons and allied groups have been that monitoring is appropriate if Glasgow Coma Scale is between 3 and 8 and there are abnormalities on CT scan, or if there is no abnormality on the CT but the patient has any two of age over 40, posturing, or has systolic blood pressure below 90 mm Hg. They set a desirable level of ICP of below 20 mm Hg and this has reinforced the role of ICP monitoring in head trauma management.
A reassessment of the effectiveness of ICP monitoring in a randomized trial reached the contrary conclusion that the information gained offers no advantage over clinical observation and imaging with CT scan. This trial was carried out by Chesnut and colleagues in developing countries and defined raised intracranial pressure at a level that has been criticized as too low (20 mm Hg). Nonetheless the study has demonstrated that the use of a clinical approach to management of raised ICP is as feasible as a program based on direct ICP measurement. This does not negate the desirability of keeping ICP controlled at some arbitrary level; it merely questions the need for direct monitoring as a guide to management.
As a practical matter, we use ICP monitoring in our unit to warn of impending deterioration from brain edema or hemorrhage if the patient cannot be effectively examined or shows poor responsiveness with evidence of mass effect on a CT scan. Although the risk of infection with a ventricular catheter is low, less than 3 percent, prolonged use may be complicated by bacterial meningitis. The catheter may be left in place for 3 to 5 days, or fewer if the clinical state and ICP are stable for 24 to 48 h. The current generation of ICP monitors employs fiberoptic strain gauges that can be inserted directly into the cerebral cortex without apparent damage.
The first step in lowering elevated ICP is to control the incidental factors that are known to raise pressure, such as hypoxia, hypercarbia, particularly hyperthermia, awkward head positions that compress the jugular veins, and high mean airway pressures from positive pressure ventilation (see the monograph by Ropper and colleagues  and Chap. 30 for further details). The avoidance of hyponatremia and serum hypoosmolarity that would allow water to enter the brain and increase its volume is accomplished by infusing only isoosmolar or hyperosmolar solutions such as normal saline. Elevations in serum osmolality as a consequence of excessive concentrations of diffusible solutes such as glucose are not useful in reducing intracranial volume because they do not create gradient for water and solutes across the cerebral vasculature. Consequently, fluids such as 5 percent dextrose in water, 0.5 normal saline, and 5 percent dextrose in 0.5 normal saline are avoided; lactated Ringer solution is permissible; normal saline, with or without added dextrose, is ideal. In a post hoc study of a cohort of severely injured patients, resuscitation with albumin was found to have a detrimental effect compared to saline (SAFE Investigators).
The basis for this class of treatments is the creation of a gradient of water concentration from the brain to the blood that reduces brain volume. Mannitol, glycerol, and urea are effective in lowering ICP by producing serum hyperosmolarity initially and then causing a diuresis that sustains this state and secondarily causes hypernatremia and hypovolemia. Hyperosmolar saline, in contrast, raises serum sodium directly and expands intravascular volume.
The effects of mannitol have been of great interest to those who treat head trauma, but the ideal plan for its use has never been established. If intracranial pressure exceeds a predetermined level, for example 20 mm Hg as recommended by aforementioned guidelines for the treatment of traumatic brain injury, mannitol 20 percent, 0.25 to 1.0 g/kg is given every 3 to 6 h to maintain serum sodium above approximately 142 mEq/L and osmolarity of 290 to 315 mOsm/L. Even if ICP monitoring is not used, an attempt may be made to maintain this level of serum osmolality for the first days after injury if contusion and brain swelling are detected on the CT scan.
Mannitol in large amounts may cause renal failure, almost always reversible, though an uncertain mechanism perhaps having to do with renal blood flow. Limited evidence suggests that this complication occurs only with the use of more than 200 g of mannitol daily.
The relative merits of hypertonic saline and mannitol are frequently discussed and have been reviewed by one of us (Ropper, 2012). Several small series comparing the agents, referenced in the review, have shown too little difference to allow a choice between the two agents. Local experience and an overall assessment of the side effects of each typically dominate practice. Hypertonic saline (concentrations of 3 percent to 23 percent) has a comparable effect to mannitol in the treatment of raised ICP and has the advantage of avoiding severe dehydration because it increases osmolarity directly rather than through diuresis. The opposite also pertains, namely that patients with poor cardiac output may be subject to congestive heart failure with hypertonic saline in high volumes. Diuretics have been used to mitigate this effect. Either agent can produce a hyperglycemic, hyperosmolar state in diabetics, particularly in the elderly and in those receiving corticosteroids.
Hypertonic saline, 3 percent, can be used in boluses of 150 mL; a 7.5 percent solution, in 75 mL boluses; and 23 percent, in volumes of approximately 30 mL. All but the lowest concentration of saline require a central venous catheter to prevent sclerosis of veins. The same levels of sodium concentration as noted for mannitol are used as a reference to guide graduated increments of sodium administration, with serum sodium higher than about 156 mEq/L infrequently providing additional reductions in ICP.
Hypocarbia, induced by hyperventilation, produces alkalosis of the CSF and cerebral vasoconstriction with a corresponding reduction in cerebral blood volume and ICP. It is effective for a limited period of time, as the pH of the spinal fluid equilibrates over hours by the elaboration of ammonium ions in the choroids plexus, allowing cerebral blood volume to return to its previous level. A single-step reduction in PCO2 typically lowers ICP for approximately 20 to 40 min. Attempts to prolong the effect of hypocarbia and the alkalosis by the intravenous administration of ammonium buffers have met with mixed success.
It has been suggested that hyperventilation may be harmful to some head-injured patients because of a reduction in cerebral blood flow, but the risk, if any, appears to be minimal, at least in adults. In children, reductions in cerebral blood flow have been demonstrated by Skippen and colleagues at even modest levels of hypocarbia and three-quarters show slight brain ischemia when PCO2 is below 25 mm Hg. For these reasons, hypocarbia is used in cases of head trauma mainly in acute circumstances and has been eschewed for chronic use. If the ICP continues to rise and brain swelling progresses despite these measures, the outlook for survival is poor. It should be mentioned that many patients, particularly children, hyperventilate spontaneously after head trauma.
Hypothermia and barbiturate anesthesia fairly consistently reduce ICP but relatively few patients respond to such measures for long and their clinical outcome is not improved. The main problem, aside from the difficulty maintaining lower body temperatures, is that rewarming induces substantial brain swelling and a return of ICP to prior levels or higher. A randomized controlled trial of cooling adult patients with severe closed head injury (Glasgow Coma Scale scores of 3 to 7) to 33°C (91.4°F) for 24 h appeared to hasten neurologic recovery and may have modestly improved outcome (Marion et al), but a larger and better-conducted study led by Clifton showed that attaining hypothermia of 33°C (91.4°F) within 8 h of injury failed to improve outcome and this approach cannot be endorsed except in special circumstances. The same lack of effect has been shown in studies with children (Hutchinson et al).
Although barbiturates lower ICP, they lower the blood pressure as well; hence they may diminish cerebral perfusion. However, an uncontrolled series by Marshall and coworkers (1979) claimed improved survival by using barbiturates even in cases where the ICP exceeded 40 mm Hg. The more definitive randomized study by Eisenberg and associates showed no benefit from barbiturate-induced anesthesia in head-injured patients, and this class of drugs has been largely abandoned except for brief, acute control of ICP while other measures are being instituted.
Several controlled studies have established that the administration of high-dose steroids does not improve the clinical outcome of severe head injury. Eclipsing smaller prior studies was the well-designed Clinical Randomization of an Antifibrinolytic in Significant Hemorrhage (CRASH) trial, conducted in more than 10,000 adults and controlled for varying degrees of injury as judged by the Glasgow Coma Scale and imaging features. The effect of the infusion of methylprednisolone 2 g, followed by 0.4 g/h for 48 h, favored survival in the untreated patients by a small but clear margin, leading to the current recommendation that steroids not be used routinely following head injury.
Blood Pressure Management
The management of posttraumatic systemic hypertension represents a difficult problem. Within hours after head injury, the sympathoadrenal response and elevation of blood pressure recedes spontaneously in a matter of a few hours or days. Unless the blood pressure elevation is extreme (greater than 180/95 mm Hg), it can be disregarded in the early stages. In animal experiments, it has been found that severe hypertension leads to increased perfusion of the brain and an augmentation of the edema surrounding contusions and hemorrhages. As mentioned earlier, edema is the main element in the genesis of brain swelling and of raised ICP in most head-injured patients (Marmarou et al). This reflects a failure of autoregulatory vascular mechanisms, with resulting transudative edema in damaged areas of the brain. The control of high blood pressure must be balanced against the risk of reducing cerebral perfusion pressure and the observation that even a brief period of mild hypotension may provoke a cycle of cerebral vasodilatation, increased cerebral blood volume, and elevated ICP in the form of plateau waves (Rosner and Becker). Observations such as these emphasize the need for immediate correction of hypotension in severely head-injured patients.
Because most therapies for elevated ICP dehydrate the patient or reduce cardiac filling pressures, thereby leading to hypotension, a middle course of avoiding both severe hypertension and any degree of hypotension seems the best compromise. In lowering high levels of blood pressure, diuretics, beta-adrenergic blocking agents, or angiotensin-converting enzyme inhibitors are generally used, rather than agents that potentially dilate the cerebral vasculature (nitroglycerin and nitroprusside, hydralazine, and some of the calcium channel blockers may present this risk). Hypotension should be corrected by vasopressor agents such as phenylephrine or norepinephrine. The precise level of blood pressure that requires treatment must be judged in the context of the ICP and the presence of plateau waves, the goal being to maintain normal cerebral perfusion pressure of 60 to 80 mm Hg, as well as the patient's previous blood pressure level; evidence of organ failure from either hyper- or hypotension, such as cardiac or renal ischemia, must also be considered.
If coma persists for more than 48 h, a nasogastric tube should be passed and fluids and nutrition should be given by this route. A basilar skull fracture, especially if there is a CSF leak, may preclude this route and compel a directly situated gastric tube. Agents that reduce gastric acid production—or the equivalent, antacids by stomach tube to keep gastric acidity at a pH above 3.5—are of value in preventing gastric hemorrhage. The prophylactic use of antiepileptic drugs, as discussed earlier, under "Posttraumatic Epilepsy," recently has been favored, but there is no evidence that delayed epileptic seizures are reduced (see Chang and Lowenstein). Only if there has been a seizure are antiepileptics given.
Restlessness is controlled by diazepam, propofol, or a similar drug, but only if careful nursing fails to quiet the patient and provide sleep for a few hours at a time. Etomidate and dexmedetomidine may be preferable for reducing agitation because they are minimally sedating. Fever is counteracted by antipyretics such as acetaminophen and, if necessary, by a cooling blanket. The use of morphine or bromocriptine to quiet episodes of vigorous extensor posturing and accompanying adrenergic activity already has been mentioned.
The need for surgical intervention during the acute posttraumatic period is decided by two factors: the clinical status of the patient and the findings on imaging. The presence of a subdural or epidural clot that is causing a shift of central brain structures calls for evacuation of the collection. The approach to these lesions has been discussed earlier. Should the elevated ICP not respond to this procedure or to the standard osmotic agents and other medical measures outlined earlier, or should the condition of the patient and vital signs then begin to deteriorate (heart rate rising, temperature rising or falling below normal, state of consciousness worsening, hemiplegia, plantar reflexes more clearly extensor), a renewed search must be undertaken for a late-occurring cerebral hemorrhage. Usually in these clinical circumstances, CT scanning will disclose a new or enlarged epidural, subdural, or intracerebral hematoma, or worsened cerebral edema. If death or severe disability is to be avoided, operation in these cases must be undertaken before the advanced signs of brainstem compression—decerebrate or decorticate posturing, hypertension, bradycardia—have appeared.
The use of decompressive craniectomy in patients with progressive and intractable traumatic brain swelling has been a subject of renewed interest, after having been practically abandoned several decades ago. Guerra and colleagues reported on 57 such patients, mostly young adults, who underwent wide frontotemporal craniectomy, unilateral in 31 and bilateral in 26. Of these, 58 percent attained surprisingly good states of rehabilitation. These authors were of the opinion that these results represented a significant improvement over the expected outcome in this particular group of patients. A similar open trial conducted by Aarabi and colleagues described 40 percent with good outcome. Similar results in children were reported by Polin and associates. The few cases with which we have been involved, mostly children operated late, have not been as encouraging.
However, these generally favorable results could not be validated in the randomized "DECRA" trial carried out by Cooper and colleagues. Decompression did indeed reduce ICP, as expected, when the intracranial contents are exposed to atmospheric pressure, but surgery did not improve outcome. The details of the operation and choice of ICP level that were chosen to trigger operation have been criticized, but this study remains the best information to date until further trials with different designs, some ongoing, clarify the issue of surgical decompression. Further trials of decompressive craniectomy after severe traumatic brain injury are being undertaken.
The treatment of the general medical diseases relating to protracted coma was outlined in Chap. 17. Each patient presents unique problems.