Prognosis in TBI

Given the large costs to brain injured patients and their families, it would be ideal to be able to prognosticate the clinical course after TBI. However, individual outcomes vary greatly, and case reports of miraculous recoveries exist. Thus it is always important to remain hopeful about outcomes.

Still, it is possible to give family members important information about outcomes. First, however, discussions should focus on the patient and family—what is their knowledge about the events that have happened? What are their expectations? What is their concept of “good” outcomes versus “bad” outcomes? What are their hopes and fears? For example, some families may believe that a “good” outcome is being able to converse with family members, whereas others may believe that a “good” outcome is the patient’s preinjury baseline. By exposing these concepts, families and medical teams can begin conversations at a common point.

Most studies on prognosis in TBI have used the Glasgow Outcome Scale (GOS), which uses the following categories:

Therefore the definitions of “severe,” “moderate,” and “good” recovery are quite broad and may not correspond to patient and/or family goals.

Some important predictive factors are listed below:

The best predictor of outcomes after TBI is the length of posttraumatic amnesia (PTA). The longer the duration of PTA, the worse outcomes generally are. Clinically, resolution of PTA occurs when patients are able to incorporate daily events into working memory. Patients with PTA of less than 24 hours generally have a quick and full recovery with few exceptions, and patients with PTA of more than 4 weeks generally have permanent deficits. However, more stringent “threshold” criteria have been used in previous textbooks to preserve hope for caregivers. These criteria state that severe disability (on GOS) is unlikely with PTA of less than 2 months, and good recovery (on GOS) is unlikely when PTA is more than 3 months.⁷ It has been shown that 21% patients with PTA between 1 and 2 months did return to work within a year, whereas only 9% of patients with PTA of more olthan 70 days were able to return to work within a year.⁸

Age is the second most important predictor of outcome. Older age is associated with worse outcomes, and previous textbooks have stated that GOS of good recovery is unlikely when age is older than 65 years.

Length of coma, defined as time from injury to follow commands, can be an important milestone as well. Generally, severe disability (GOS) is likely when length of coma is more than 2 weeks. Good recovery (GOS) is unlikely when length of coma is more than 4 weeks. For more information, please see the section on Disorders of Consciousness.

The best Glasgow Coma Scale (GCS) score within 24 hours correlates with outcomes, with lower scores associated with worse outcomes, and with the best motor response subset as the most accurate predictor of outcomes. Unfortunately, there are no “threshold” criteria because patients with severe injury can have good recovery. Indeed, recent reviews of GCS in outcomes of TBI suggest that other measures, such as age and pupillary response, improve its predictive capacity.⁹

Neuroimaging has become increasingly important in predicting outcome. It has been suggested that an adequate threshold marker is the presence of bilateral brainstem lesions; good recovery (GOS) is unlikely when bilateral brainstem lesions are present on an early MRI. CT scan findings of subarachnoid hemorrhage, midline shift, extradural hematoma, bilateral injury, or transventricular injury correlate with worse outcomes, though a full range of outcomes is still possible with these injuries.⁷

Research continues in this field to delineate other possible markers of prognosis and outcome. Biomarkers such as S100, genetic markers such as APOE 4 status, the ratio of total brain volume to third ventricle size, T2 gradient echo MRI lesions, PET scanning, diffusion tensor imaging, magnetic resonance spectroscopy, and somatosensory evoked potential responses are currently under study and may provide important prognostic information in the future.¹⁰

Given the uncertainty surrounding various predictors, several authors are currently trying to mathematically model outcomes of TBI using predictive variables.^11–13 One such model, based on a large population study, is available at http://www.tbi-impact.org/. It must be noted that these models are still undergoing refinement, though they may also provide important prognostic information in the future.

Introduction

Much basic science research has been done on the mechanisms of neural recovery, particularly in stroke literature. Exact comparisons may be difficult for TBI because the nature of the injury is often diffuse but it seems likely that the basic mechanisms of recovery are similar. Trauma disrupts neuronal connections including the connectivity and organization of networks mediating executive functions and sensory and motor responses.¹⁴ In addition to the focal damage of TBI, dysfunction can occur in areas anatomically removed, likely because of diaschisis involving tissue hypometabolism, neurovascular uncoupling, and aberrant neurotransmission.¹⁵ In addition to local repair, neuroplasticity is thought to be a crucial feature of recovery from TBI with research from as far back as the 1960s demonstrating how brain structure and chemistry can be affected by environmental manipulation.¹⁶

Overview

Recovery of neurological function is dynamic and multifaceted but often limited partly because of the inability for dramatic anatomic reorganization in the brain following lesions. Three distinct phases of recovery have been described. The first involved activation of cell repair and reversal of diaschisis; the second, plasticity of existing neuronal pathways to produce a functional change; the third, neuroanatomic plasticity to produce nascent connections. Where damage is focal, recovery primarily engages ipsilateral brain tissue, but with more severe TBI it is likely that contralateral brain regions are also recruited.¹⁶ The recovery process begins immediately and extends over the course of many months to years.

Therapeutic Targets

To date there are no currently approved pharmacologic agents that help restore neural function lost in TBI. With a better understanding of inhibitory mechanisms that block regenerations and processes that enhance plasticity, newer therapeutic targets will become available.

One area of interest is growth promoters and inhibitors. When administered within days of the insult, functional recovery is promoted by growth factors through stimulation of neuronal plasticity. Several agents are currently in clinical trials and are diverse in nature, including erythropoietin, granulocyte colony stimulating factor, statin drugs, and phosphodiesterase-5 inhibitors.¹⁵ By contrast, growth inhibitors impair recovery with particular insights being gained into the effects of growth cone extension. Myelin-associated proteins, including Nogo-A, oligodendrocyte myelin glycoprotein, and myelin-associated glycoprotein (MAG), activate the inhibitory Nogo-66 receptor to prevent axonal sprouting.¹⁷ Intravenous or intrathecal delivery of antibodies or peptides aimed at either the Nogo protein or receptor dramatically enhance functional recovery after neural injury in animal models.

Finally, therapy may in the future focus on brain activation with early studies showing that amphetamine and possibly L-dopa enhanced recovery of function in animal models. The suggestion was that adrenergic and cholinergic neurotransmitter systems are critical to the recovery process, but human trials have had limited results.¹⁸ Brain tissue can also be activated by physical and environmental means, for example, repetitive physical activity or sensory experiences. As part of environmental stimulation, social interaction is felt to be important in that it induces multiple biologic effects on the brain.

While many questions remain about the mechanisms of recovery in TBI, the current sophisticated molecular tools available for researchers continue to yield insights and hopefully will lead to promising new therapies.¹⁵

Disorders of Consciousness

Up to 15% of patients with severe TBI are unable to respond consistently to commands at 4 weeks after injury, and the majority of these will have impairments of consciousness.¹⁹ There has been much confusion surrounding the different states of consciousness. By definition:

Coma—unarousable neurobehavioral unresponsiveness. Patients remain with eyes closed, without evidence of eye opening either spontaneously or in response to external stimulation. They do not follow commands, do not demonstrate goal-directed or volitional behavior, do not verbalize or mouth words, and cannot sustain visual pursuit movements beyond a 45-degree arc. Neurobehavioral signs and symptoms secondary to pharmacologic treatment with agents such as paralytic or sedative drugs must be excluded.

Vegetative state (VS)—arousable neurobehavioral unresponsiveness. Patients have periods of eye opening, either spontaneously or in response to external stimuli; subcortical responses to external stimuli (including posturing, tachycardia, diaphoresis with pain); return of autonomic functions (sleep-wake cycle, normalization of respiratory and digestive functions); roving eye movements without concomitant visual tracking ability.²⁰

Minimally conscious state (MCS)—a condition of severely altered consciousness in which the patient demonstrates minimal but definite behavioral evidence of self- or environmental awareness. The Aspen workgroup proposed that one or more of the following should be present: (1) following simple commands; (2) manipulation of objects; (3) gestural or verbal yes/no responses; (4) intelligible verbalization; or (5) stereotypical movements that occur in a meaningful relationship to the eliciting stimulus and are not attributable to reflexive activity. These responses must be reproducible or occur on a sustained basis.²¹

Conscious—reliable and consistent demonstration of interactive communication or functional use of objects.

Overall, there has been some controversy over the terms “persistent” and “permanent” vegetative state. The American Academy of Neurology has suggested that persistent vegetative state be defined as more than 1 month, while permanent vegetative state as more than 3 months for nontraumatic injury and more than 12 months for traumatic injury. However, there have been case reports of patients regaining consciousness more than 1 year out from their injury, and in one case series, 52% of patients who were vegetative at 3 months recovered consciousness by 12 months postinjury. Because of these issues, some experts have suggested eliminating the terms “persistent” and “permanent” altogether, and to use a probability for outcomes—for example, it is highly likely that patients who recover consciousness more than 1 year postinjury are severely disabled (unable to be independent for a 24-hour period of time).

Amantadine, bromocriptine, amphetamine, and methylphenidate and other neurostimulants have been proposed to help with arousal and vigilance.²² There has been a case report of a patient in MCS who regained full consciousness after treatment with amantadine who declined when the medication was withdrawn.¹⁰ However, Cochrane Reviews on monoaminergic agonists and excitatory amino acid inhibitors have not found good randomized trials to review.^23,24 Similarly, another recent Cochrane Review determined that there have not been any good studies to support the use of sensory stimulation in MCS or VS patients.²⁵ Unfortunately, most research surrounding treatment of disorders of consciousness in brain injury is poor, involving case reports or small numbers of patients without controls. More research is needed to further clarify treatment algorithms in the future.

Motor Recovery

As with everything in brain injury, motor recovery is highly individualized and varies from person to person. Though it seems that the majority of motor recovery happens in the first 6 months after injury, it is possible for patients to continue gaining function after that. Swaine and coworkers studied a small population of patients with severe TBI (GCS ≤ 8) for 6 weeks after injury, and demonstrated that, while the percentage of patients with basic primitive reflexes (asymmetric tonic neck reflex, flexor withdrawal) stayed relatively stable, equilibrium reactions, protective reactions, sitting reactions, and simple motor skills slowly increased over time. The sequence of recovery—reactions in sitting appeared before those in kneeling and before those in standing. Recovery of equilibrium reactions tended to occur before development of protective reactions. Similarly, with motor skills, rolling ability recovered before ability to sit unsupported, followed by the ability to stand, then walk independently, then climb stairs, then hop on one foot, then the ability to run. Surprisingly, some simple motor skills (rolling, sitting, even walking unsupported) seemed to recover faster than equilibrium. Also surprisingly, patients did not necessarily lose primitive reflexes before recovering simple motor tasks. Of note, in this case series, 60% of patients by 6 weeks could sit independently and 56% of patients could walk without assistance.²⁶ Watson and associates have suggested a path-tree analysis instead of a simple linear progression of motor skills and demonstrated that while there is an underlying pattern of recovery, there are many routes that a patient can follow.²⁷

Motor Trephine Syndrome

Following craniectomy, patients may suffer delayed or insidious headache, fatigue, and mood disturbance. A few may have delayed contralateral monoparesis that is reversible with cranioplasty—the motor trephine syndrome. In one retrospective review, up to 26% of patients with decompressive craniectomy developed delayed contralateral upper extremity weakness that improved with cranioplasty, with full recovery 1 month after cranioplasty. Weakness developed an average of 4.5 months later, was predominantly distal upper extremity proximal, causing difficulty with grip and extension of metacarpals, and was slowly progressive. There was no sensory or language impairment. These patients more often had blossoming of cerebral contusions after craniectomy and cerebrospinal fluid (CSF) flow disturbances (hygromas and hydrocephalus), and had a longer time to cranioplasty than patients without motor trephine syndrome. Mechanistically, it has been proposed that motor trephine syndrome is caused by disturbances of CSF flow, leading to edema in brain parenchyma. Cranioplasty allows egress of CSF, which resolves brain edema and restores function.²⁸

Evaluation of Cognitive Dysfunction

Initial evaluation of cognitive dysfunction, as with any neuropsychological or behavioral dysfunction, begins with medical evaluation. Infection, electrolyte imbalances, hypoxia or hypercapnia, endocrine disturbances, new neurological lesions (strokes or hydrocephalus), medications, medication withdrawal, sleep disturbances, seizures, and reactive mood disorders can all impair cognition in hospital patients. Of course, documentation of baseline cognitive abilities of the patient is important. After correction for these factors, formalized neuropsychological testing should be considered to fully evaluate the cognitive domains that are affected—language, visuospatial, attention/concentration (which includes arousal, selective attention, effort, and resource allocation), learning and memory, executive functioning, sensory perceptual, and psychomotor-visuomotor functioning. Neuropsychiatric testing can also help with determination of premorbid intelligence if previous data do not exist. Ideally, pharmacologic and nonpharmacologic treatment could then be targeted to deficits, though research is still preliminary in these fields.

Treatment of Cognitive Dysfunction

Overall, there have been very few class I studies to guide treatment of cognitive dysfunction in brain injury.³² However, some possible targets and treatments are outlined below (Table 342-1).

Methylphenidate: Recently, a Cochrane review concluded that there was not enough evidence to warrant using a monoaminergic agonist to improve recovery after TBI because none of the studies done fulfilled inclusion criteria.²⁴ Despite this, recent reviews have mentioned that the several small scale randomized controlled trials (RCTs) involving methylphenidate are among the strongest studies in pharmacologic therapy in brain injury. Overall, these reviews suggest that methylphenidate treatment leads to significant improvements for speed of processing, and certain areas of attention and concentration, such as attentiveness. It has also been suggested that those with slower baseline information processing demonstrate a greater drug response. Thus there is relatively strong evidence that methylphenidate improves speed of processing and attention in traumatic brain injury patients.^33,34 Dosages that have been used in the literature include 20 mg or 0.3 mg/kg.^10,35

Donepezil and other cholinergic medications: Donepezil, galantamine, and rivastigmine, all cholinergic agonists, have not been as rigorously studied as methylphenidate. There has been one small RCT and a few small case series that concluded an increase in sustained attention and short-term memory in donepezil up to 10 mg/day. One open label study of galantamine, donepezil, and rivastigmine demonstrated some positive response in attention and vigilance, but no difference in side effects. A single large rivastigmine trial failed to show benefit in all groups; however, those with more significant memory impairments had a trend toward improvement.³⁶ Overall, there has been moderate evidence of the use of cholinergic medications to enhance attention and concentration.³⁷

Amantadine: Amantadine is a dopamine receptor agonist and N-methyl-D-aspartate (NMDA) NMDA receptor antagonist. Case studies have suggested positive response with short-term memory, attention, planning, impulsivity, and disinhibition.³⁸ There has been one small double-blind, crossover RCT that did not demonstrate significant difference in rate of cognitive improvement.³⁹ Despite conflicting information, recent reviews in 2008 suggest “at doses of 200 to 400 mg/day, amantadine appears to safely improve arousal and cognition in patients with TBI,” though they warn that more studies are necessary.^37,40 Small case reports have suggested some benefit for other dopaminergic agents, such as bromocriptine, selegiline, and levodopa, but the data are still very limited with regard to these agents.

Sertraline and other SSRIs: It is well known that depression may confound cognitive impairment in TBI. Recent consensus opinion encourages the use of antidepressants in TBI, though notes that there may not be a strong effect on cognition. They suggest the use of sertraline as a first-line agent because it does not have as many side effects. One small, nonrandomized study without control suggested that fluoxetine may have improved attentional task and working memory in chronic TBI patients.^35,41

Modafinil and atomoxetine: Although there have not been any human studies on the effect of atomoxetine on cognition in TBI patients at this time, animal studies have been promising.⁴² Modafinil has been studied for underarousal and fatigue in TBI, but has not been studied in cognition.^43,44 Still, one recent review suggests that, given their effect on the dorsolateral prefrontal cortex, and mechanisms of action, atomoxetine and modafinil may be useful for attention and arousal in TBI patients.³⁵

Warnings: Although one class I study has suggested that valproate does not impair or improve cognitive function, many other medications used in brain injury may inhibit or impair cognitive function. For example, although no studies have been done on topiramate and TBI, some authors suggest that topiramate use may cause neuropsychological problems in patients who have underlying cognitive and behavioral problems.⁴⁵ The use of neuroleptics, such as haloperidol, is controversial in brain-injured patients because they may increase the length of PTA and have a negative impact on cognitive recovery.⁴⁶ In general, practitioners should be wary of neuroactive medications in the brain-injured patient, and seek to eliminate medications with cognitive side effects.

Definition

Following TBI, a subset of patients experience dysautonomia, a syndrome characterized by severe paroxysmal changes in heart rate, respirations, blood pressure, temperature regulation (including diaphoresis), and muscle tone.⁴⁹ Dysautonomia has been called various other names, including autonomic dysfunction syndrome, sympathetic storming, hypothalamic-midbrain dysregulation syndrome, acute midbrain syndrome, and diencephalic epilepsy.

Incidence and Effects

Dysautonomia in particular is associated with severe TBI. Other associations include severe diffuse axonal injury, preadmission hypoxia, younger age at time of injury, and possibly brainstem injury.⁵⁰ In the intensive care setting, the incidence of dysautonomia ranges from 8% to 33%, falling to 5% in the rehabilitation setting. This decreased incidence is consistent with the observation that over time paroxysmal autonomic overactivity gradually settles, coinciding with neurological recovery. TBI patients with dysautonomia, however, are more likely to experience significant long-term disability, both cognitive and physical, and have longer hospital and rehabilitation stays; length of ICU stay is similar to the nondysautonomic group.^50,51 A discrepancy in reported incidence is evidence of variability in criteria used in research and highlights the need for standardization in the literature.

Mechanisms

Understanding of the pathophysiology of dysautonomia in TBI remains problematic. Original thinking on the subject assumed an epileptogenic source, hence the popular moniker diencephalic seizures. Multiple attempts to identify epileptic discharges or treat epilepsy have yielded negative results. More recently this has been superseded by a range of disconnection theories. Autonomic control exists at several levels of the nervous system and disconnection theories suggest that dysautonomia is a result of the liberation of excitatory centers from higher central control. Debate exists as to whether the excitatory centers are located in the upper brainstem and diencephalon or in the spinal cord. More conventional theories suggest that the upper brainstem and diencephalon drive the autonomic paroxysms. An alternative theory suggests that these centers are inhibitory on spinal cord processes and damage to the centers or their connection to the cord results in inappropriate spinal cord autonomic activity.

Evidence suggests that the disconnection syndrome can result from structural and/or functional disconnection. Structural damage results from diffuse and focal injury sustained to the central nervous system in TBI patients. In contrast, functional disconnections may occur because of neurotransmitter abnormalities or through alterations of the functional environment (e.g., raised intracranial pressure [ICP]).⁵²

Anecdotally, various afferent stimuli have been identified as dysautonomic triggers including endotracheal tube suctioning, Foley catheter manipulation, passive movements (e.g., repositioning patient), and muscle stretch.⁴⁹

Symptoms

Identifying dysautonomia in the intensive care unit (ICU) or rehabilitation setting should be based on the operational definition of paroxysmal increases in at least five of the following seven clinical features: heart rate greater than 120 beats/min, respiratory rate greater than 30, temperature, systolic blood pressure greater than 160, sweating/flushing/piloerection, muscle tone, and posturing (e.g., decerebrate or decorticate). By definition, several of these components will occur simultaneously and typically to a marked degree.^53,54 The higher metabolic demand in the posturing dysautonomic patient, combined with postinjury gastrointestinal abnormalities results in an estimated 25% loss of body weight because of a highly catabolic state. This malnutrition increases the risk of infections and the development of critical illness neuropathy. In addition, dystonic posturing in the setting of weight loss increases the risk of developing pressure injury and contracture.⁵⁵

Treatment

The first step of any treatment paradigm should be to minimize noxious stimuli in patients at risk for or with documented dysautonomia. This includes aggressively identifying and treating pain, decubitus ulcers, infections, constipation, minor and undiagnosed injuries, and heterotopic ossification. For any patient that develops symptoms consistent with dysautonomia, the differential should remain wide and appropriate evaluation for other conditions that might produce similar findings including neuroleptic malignant syndrome, posttraumatic epilepsy, and pulmonary embolism should be conducted.

A wide range of medications has been used to treat patients with this disorder and have anecdotal support in the literature. These include morphine, α-blockers (predominantly clonidine), β-blockers (predominantly propranolol and metoprolol), anticonvulsants (valproic acid and phenobarbital), dopamine agonists (in particular bromocriptine), and benzodiazepines.⁵⁶ Much of the therapeutic approach has been directed at reducing efferent activity and more recent literature has suggested benefits from both intrathecal baclofen and gabapentin administration. Both are GABA analogues and it is possible that both act to normalize modulation of afferent stimuli in dysautonomic patients by enabling inhibitory pathways within the spinal cord, thus preventing the development of the efferent arm of the syndrome.⁵⁵

Each particular patient is likely to require a combination of drugs to ameliorate symptoms but it seems reasonable to initiate and uptitrate gabapentin therapy (in light of the benign side effect profile) as an initial step before other agents.

Treatment of Agitation

Similarly to cognitive dysfunction, treatment of agitation begins with a broad differential diagnosis of delirium. This differential should include medical issues such as pain, infection, hypoglycemia/hyperglycemia or other electrolyte imbalances, hypoxia, epilepsy, endocrine issues, medications, and medication withdrawal. Once those are ruled out, the diagnosis of posttraumatic agitation can be made.

Once other causes of increased psychomotor activity and confusion are ruled out, the first step in treating agitation is limiting environmental stimuli, especially at night. Though difficult in an acute care setting, limiting ambient noise, light, assessment of vital signs, and around-the-clock medication administration can help reset the circadian rhythm and thereby limit agitation. Limiting use of lines, tubes, and restraints, if possible, can also help reduce unwanted psychomotor activity. Abdominal binders and IV protectors can be useful in preventing patient-induced discontinuation of lines and tubes.

Once environmental causes have been treated, the option exists to use pharmacologic treatment. Though a recent Cochrane review of pharmacologic treatment of aggression or agitation in TBI noted that most studies were case studies, and therefore did not meet inclusion criteria, some possible treatments are listed below.

Medications to Avoid: Antipsychotics

Recently, there has been increasing concern with the use of antipsychotic medications in TBI patients. TBI experts often avoid these medications more than nonspecialized physicians.⁵⁸ While there have been case studies suggesting decreases in agitation with medication administration, animal studies with typical antipsychotics have had slower motor recoveries. Typical antipsychotics such as haloperidol have been shown to correlate with longer periods of posttraumatic amnesia in human studies, although some speculate that this may be caused by increased usage of antipsychotic in more severe injuries. Patients who are tapered off chronic typical antipsychotic medications have improved neuropsychiatric studies. While atypical antipsychotics such as olanzapine, clozapine, and risperidone have not had similar effects in animal models, recent work has suggested that high doses of risperidone may have similar cognitive effects in humans as haloperidol.

Medications to Avoid: Benzodiazepines

Benzodiazepine use in TBI has also caused concern in TBI experts. Though effective for treatment of agitation, animal models have suggested that use of benzodiazepines both acutely or chronically can lead to further cognitive impairment. Also, normal patients, when weaned off chronic benzodiazepines for insomnia, have been noted to have neuropsychological deficits when compared with controls. Therefore, several experts have suggested that benzodiazepines may be detrimental in patients with TBI and should not be prescribed in these populations.

Posttraumatic Epilepsy