Differential Diagnosis and Functional Outcome
Long-term Prognosis and Functional Outcome
Prognosis and outcome for patients with disorders of consciousness have been analyzed in three main areas: mortality, recovery of consciousness, and functional recovery. Level of consciousness and length of time postinjury are key predictors of prognosis. However, the challenge of detecting consciousness and discerning level of awareness often confounds prognostic formulations. Evidence with respect to diagnostic inaccuracy and ambiguities depending on methods of assessment illustrate these challenges. Outcome studies have historically been compromised by poorly described or inaccurate diagnostic characterization, further compromising the existing base of knowledge concerning disorders of consciousness. Recent investigations have begun to resolve these problems, shedding new light on assessment, treatment, and long-term outcome.
Differential Diagnosis. In long-term outcome studies of patients with disorders of consciousness, the etiology of brain injury has usually been dichotomized as traumatic brain injury (TBI) versus nontraumatic brain injury (non-TBI), including anoxic brain injury; stroke; and infectious, toxic, and metabolic disorders. Although further subcategorization might inform prognosis, prognostic information regarding specific etiologies of non-TBI or subtypes of TBI is lacking.
Over the years, several terms have been used to characterize altered states of consciousness, but the most widely accepted terms at present are coma, vegetative state, and minimally conscious state. Acute confusional state, posttraumatic confusional state, or simply confusional state have been proposed to describe the condition of those who have recently emerged from a minimally conscious state. This condition may also be considered a disorder of consciousness, albeit at a more established level of awareness and purposeful responsiveness.
Coma is a state of pathologic unconsciousness in which the eyes remain closed and the patient cannot be aroused. The defining feature of coma is the complete loss of spontaneous or stimulus-induced arousal. On examination, there is no evidence of purposeful motor activity, no response to command, and no indication of receptive or expressive language ability. Coma is a self-limiting state that typically resolves within 2 to 4 weeks in those who survive the initial injury.
The vegetative state is a condition in which there is no sign of consciousness, awareness of self or environment, with complete absence of purposeful behavior, although the capacity for spontaneous or stimulus-induced arousal is preserved. The vegetative state typically follows a period of coma, but may also occur in those with developmental malformations or may represent late stages of progressive degenerative or metabolic disorders. The term itself is often perceived by the public as having pejorative connotations. This has led to occasional efforts to abandon the term; in a proposal by the European Task Force on Disorders of Consciousness, the term "unresponsive wakefulness" was recently suggested to replace vegetative state.
The American Academy of Neurology (AAN) recommended that the term "persistent vegetative state" be used to describe a vegetative state when it lasts for at least 1 month following TBI or non-TBI. Because of the ambiguity of the term persistent in implying prognosis, the Aspen Neurobehavioral Workgroup recommended that the term "persistent vegetative state" be abandoned in favor of accompanying the diagnosis of a vegetative state with a description of the injury and the length of time since onset, as both of these factors provide prognostic information. The AAN practice guideline also recommended that the term "permanent vegetative state" be applied 3 months after non-TBI, and 12 months following TBI, implying that further recovery of consciousness is highly unlikely after these intervals. More recent evidence described below suggests the term "permanent" should also be abandoned, especially in these time frames.
A minimally conscious state is a condition of severely altered consciousness in which there is minimal, but definite behavioral evidence of self or environmental awareness. A minimally conscious state usually represents a transitional state reflecting improvement in consciousness from coma or a vegetative state, or progressive decline as in neurodegenerative disease (e.g., Alzheimer disease). The diagnosis of a minimally conscious state is based on clearly discernible and reproducible evidence of at least one of the following behaviors: simple command following, intelligible verbalization, recognizable verbal or gestural "yes/no" responses (without regard to accuracy), or movements or emotional responses triggered by relevant environmental stimuli that cannot be attributed to reflexive activity. In making the diagnosis of a minimally conscious state, it is important to consider the frequency and complexity of the behavior observed, factors that challenge accurate diagnosis. When there is evidence of rudimentary cognition only (e.g., visual tracking), at the lower end of the minimally conscious state spectrum, the diagnosis of a minimally conscious state requires serial reassessment to confirm that these behaviors are contingently related to meaningful environmental stimuli. At the higher end of the minimally conscious state spectrum, patients who demonstrate reproducible evidence of complex cognitively mediated behavior (e.g., following commands with movements that are extremely unlikely to have occurred by chance, responses signaling yes/no intent or intelligible verbalization) clearly confirm at least a minimally conscious state level, even if very inconsistent. Temporal fluctuations in the behaviors that distinguish a vegetative state and minimally conscious state are not unusual and may lead to diagnostic instability. Some centers have adopted local norms for diagnostic coding that require the behavioral criteria to be met over consecutive examinations before a diagnosis is established.
Emergence from a minimally conscious state has been defined by the recovery of reliable and consistent interactive communication or functional object use. Communication may be demonstrated through verbal responses, gestural responses or using augmentative devices. Patients who emerge from a minimally conscious state may be considered in a confusional state, or posttraumatic confusional state in the case of TBI, a condition that is characterized by significant impairments in attention and anterograde amnesia (labeled posttraumatic amnesia after TBI). Aphasia or apraxia may confound the accurate diagnosis of the transition from a minimally conscious state to a confusional level of recovery. It has also been suggested that the accuracy of diagnosis of emergence from a minimally conscious state to a confusional state may be compromised by the requirement of reliable and consistent interactive communication. Studies by Nakase-Richardson and colleagues raised the question of whether the communication criterion for emergence from a minimally conscious state (i.e., consistently accurate responses to yes/no questions) is too stringent, after providing evidence that some patients who had reportedly emerged from a minimally conscious state but remained confused, failed to accurately respond to simple yes/no questions.
Bedside examination is the mainstay of clinical assessment, but specialized behavioral and neuroradiologic protocols have been developed in an attempt to provide a more specific means of monitoring recovery across the subacute and postacute periods and to improve outcome prediction. There are no standardized evaluation procedures for the clinical bedside examination of patients with disorders of consciousness. Clinicians rely on systematic evaluations of arousal and behavioral responses to various forms of stimulation. Diagnostic errors arise because of a misinterpretation of behaviors or examinations that are inadequate to detect low level, inconsistent responsiveness, or because more specific sensory (e.g., blindness), motor (e.g., paralysis), or cognitive deficits (e.g., aphasia, apraxia) prevent the patient from demonstrating consciousness through specific assessment tasks. The bedside neurologic examination of patients with impaired consciousness should focus on two general areas, assessment of the integrity of the central nervous system, particularly brainstem pathways (e.g., pupillary responses, ocular movements, oculovestibular reflexes, breathing patterns) and the presence of higher-level cortical functions (e.g., purposeful, intentional behaviors in relation to external or internal stimuli). The examiner's task is to systematically elicit and distinguish behaviors that are nonspecific, reflexive, or automatic, from those that are stimulus specific and mediated by integrated cortical activity representing some level or awareness or purposeful intent.
Cognitive awareness or conscious intent may be difficult to interpret when responses are extremely inconsistent or simple, and arousal and responsiveness may fluctuate considerably over time. Single bedside examinations are usually inadequate to conclusively diagnose level of consciousness. Repeated assessments and use of standardized evaluations (see below) are often necessary to assure diagnostic accuracy.
Standardized Rating Scales. Whereas the Glasgow Coma Scale was the only standardized scale for disorders of consciousness in the recent past and remains a dominant metric for acute assessment and prognostication, neurorehabilitation specialists have recognized the need for reliable assessment instruments that can detect more subtle, but potentially important clinical findings to guide diagnosis and prognosis. A variety of standardized neurobehavioral assessment scales for disorders of consciousness have been developed over the past 20 years with varying validity, reliability, and diagnostic and prognostic utility. The Disorders of Consciousness Task Force of the American Congress of Rehabilitation Medicine completed a systematic review of 13 behavioral assessment scales for disorders of consciousness. Although the results of this review yielded insufficient evidence to support the use of any of these measures for use in outcome prediction, one scale, the Coma Recovery Scale-Revised (CRS-R), was recommended by the Task Force with "minor reservations" for use in individuals with disorders of consciousness. It was also recently selected as the measure of choice for assessment of recovery of consciousness in TBI research by the Traumatic Brain Injury Common Data Elements (CDE) Outcomes Workgroup, cosponsored by the National Institute of Neurological Disorders and Stroke (NINDS), Defense Centers of Excellence for Psychological Health and Traumatic Brain Injury, U.S. Department of Veterans Affairs, and the National Institute on Disability and Rehabilitation Research. The CRS-R is comprised of six subscales addressing auditory, visual, motor, oromotor/verbal, communication, and arousal functions with subscale items hierarchically arranged, corresponding to brainstem, subcortical, and cortically mediated functions. The scale has been used to investigate diagnostic accuracy, the relationship between behavioral and functional imaging markers of consciousness, outcome prediction, and treatment effectiveness. Additional information is available on the NINDS TBI CDE website (http://www.commondataelements.ninds.nih.gov/TBI.aspx).
Individualized Behavioral Assessment. The rigorous attention to methodological consistency afforded by standardized measures does not always allow case-specific questions to be addressed. For example, standardized procedures that restrict the number of observations may not be able to differentiate between random movements and inconsistent, low-frequency movements to command. Whyte and DiPasquale developed an approach termed "Individualized Quantitative Behavioral Assessment" (IQBA) to address this problem. The IQBA applies principles of single-subject research design to assess the cognitive and behavioral capacities of individuals with marked limitations in responsiveness. Using this technique, clinical questions are individually tailored, stimuli and response criteria operationally defined and behavioral frequencies analyzed statistically to determine whether the behavior exceeds the rate predicted by chance. This approach has been successfully employed to investigate command following, visual discrimination, and communication ability.
Estimates of misdiagnosis among patients with disorders of consciousness have been very high, ranging from 37 to 43% in some studies. It was thought that accuracy should be higher, after diagnostic criteria for a minimally conscious state were established in 2002; however, a more recent study on diagnostic accuracy showed otherwise. Using a standardized assessment instrument the CRS-R as the gold standard, the authors found that 41% of the diagnoses established by expert team consensus were incorrect, 89% of which were in the false-negative direction (i.e., conscious, but believed to be unconscious). This rate is equivalent to the rates reported before the criteria for a minimally conscious state were published.
Many factors have been implicated as causes of diagnostic error in disorders of consciousness, including patient, examiner, and environmental influences. There are several common problems that can lead to inaccurate diagnosis in patients with impaired consciousness. These include attributing purposeful intent to behaviors that are nonpurposeful, reflexive, or generalized responses; inadequate evaluation to detect conscious behavior, such as insufficient sampling time, inadequate arousal, inappropriate choice of stimuli; failure to account for sensory, motor, or cognitive impairments that confound assessment results; over- or underconsideration of family or other's observations of purposeful behavior (i.e., failure to recognize that family may be first to observe signs of consciousness or to overattribute purposeful intent); and overattributing intent to simple, cortically mediated behaviors of uncertain cognitive significance, such as simple isolated limb movements.
To improve diagnostic accuracy and provide another means to assess cognitive awareness and intent, investigational procedures have been developed using functional magnetic resonance imaging and recently electrophysiologic techniques (i.e., evoked response potentials, quantitative electroencephalogram [EEG]) to assess brain activity during different activation protocols. The protocols involve passive paradigms, such as presentation of speech or complex visual stimuli, or active paradigms, such as imagining a sporting activity or navigating through a familiar environment. Consciousness is inferred when the pattern of brain activity is analogous to controls. Although there is correspondence with behavioral evaluations much of the time, there is sometimes incongruity between findings on behavioral examinations and functional imaging. Large studies assessing the sensitivity and specificity of these assessment techniques, in comparison to behavioral evaluation, have not been published.
Outcome Studies of Disorders of Consciousness
Mortality. Mortality in patients with disorders of consciousness must be evaluated in the context of multiple factors including type of injury, injury severity, level of consciousness, timing of presentation, and duration of follow-up. Mortality rates had been considered relatively high for patients with prolonged disorders of consciousness, largely based on the Multi-Society Task Force report on persistent vegetative state (82% at 3 y and 95% at 5 y for patients in a vegetative state at least 1 mo). In a more recent study of 51 patients followed for 5 years postinjury who were either at a vegetative state or a minimally conscious state level after TBI or non-TBI at admission into an intensive care unit, mortality was lower for those admitted in a minimally conscious state (36%) compared with those admitted in a vegetative state (75%). As might be expected, mortality rates are lower for patients with disorders of consciousness who survive to rehabilitation admission. In a recent study of 50 patients admitted to rehabilitation who were still in a vegetative state after TBI or non-TBI at 6 months postinjury, mortality was 42% at a mean follow-up of 25.7 months after onset. Mortality was much lower, only 8% at a mean of 2.1 years postdischarge for a series 337 patients with disorders of consciousness, unable to follow commands on admission to 16 rehabilitation centers of the NIDRR TBI Model Systems Program. The lower mortality in this study compared with the previous study is understandable because the sample was restricted to TBI and most patients had less severe injuries with disorders of consciousness of shorter duration than 6 months. It is important to note that mortality rates rarely account for cases in which artificial nutrition and hydration were withdrawn. One exception is a recently published Canadian study that reported an overall mortality rate of 32% in six level I trauma centers. The authors found that 70% of the deaths reported were attributable to withdrawal of life-sustaining therapy, half of which occurred within the first 72 hours of injury. These data raise serious questions about basing major care decisions on prognostic indicators acquired acutely after injury.
Recovery of Consciousness. In 1994, the Multi-Society Task Force on PVS (persistent vegetative state) summarized prognosis for recovery of consciousness after a vegetative state of at least 1 month in a meta-analysis including 434 adults and 106 children with TBI, and 169 adults and 45 children with non-TBI—primarily anoxic brain injury and stroke. Prognosis for recovery was substantially better for victims of TBI than those with non-TBI. Of adults with TBI who were unconscious at least 1 month, 33% recovered consciousness by 3 months postinjury, 46% by 6 months, and 52% by 1 year. Approximately 35% of patients with TBI who were still unconsciousness at 3 months, regained consciousness by one year; if still unconscious for 6 months, 16% regained consciousness by 1 year. Of those adults with non-TBI unconscious for 1 month, only 11% recovered consciousness by 3 months and 15% by 6 months. No person with non-TBI regained consciousness after 6 months postinjury. Prognosis in children was only slightly more favorable.
The Task Force concluded that prognosis for recovery of consciousness was very poor 12 months after TBI and 3 months after non-TBI for both adults and children and suggested the term "permanent vegetative state" for those unconscious beyond these intervals after injury. However, as Childs and Mercer pointed out, there was insufficient evidence in the Task Force data to predict the frequency of late improvement. In fact, using the limited number of cases followed beyond 12 months in the Task Force report, they recalculated that the incidence of regaining consciousness after 12 months in that series of patients was 14%. Greater prospects for late recovery of consciousness are supported by a subsequent report of 50 patients with TBI and non-TBI (hemorrhagic stroke or anoxic brain injury) of even greater severity, who were in a vegetative state for at least 6 months and followed for a mean of 25.7 months after injury. At latest follow-up, 24% recovered to a minimally conscious state level, 20% of them after 12 months postinjury. It is noteworthy that of the 10 patients who recovered beyond a vegetative state level after 12 months, 4 were patients with non-TBI (3 with anoxia and 1 with hemorrhagic stroke), well beyond the 3-month designation for "permanence" by the Multi-Society Task Force. Six of these patients went on to emerge from a minimally conscious state to a higher level of consciousness (four with TBI, two with non-TBI), all remaining at a severe to extremely severe level on the Disability Rating Scale (DRS) at latest follow-up. Another report examined long-term recovery of patients with TBI or non-TBI in a vegetative state (n = 12) or a minimally conscious state (n = 39) over 1 to 5 years postinjury. Fischer and colleagues found that although none of their patients in a vegetative state recovered consciousness (10 of 12 with non-TBI), one-third of those in a minimally conscious state emerged to a higher level of consciousness after 1 year, falling within the severe disability category of the Glasgow Outcome Scale.
Nakase-Richardson et al reported on recovery of consciousness in a large cohort enrolled in the National Institute on Disability and Rehabilitation Research- (NIDRR-) sponsored TBI Model Systems. All patients had prolonged disturbance in consciousness on admission to rehabilitation, defined as a GCS motor subscale score < 6 and failure to follow commands on two consecutive days during the acute care stay. Among the 396 patients followed, 68% regained consciousness during inpatient rehabilitation. The authors also reported on 108 patients who had consistent follow-up over a 5-year period. Of these, 71 (66%) recovered command following during the course of inpatient rehabilitation. Within the subgroup of patients who failed to recover command following by rehab discharge, 54% did so by 1 year postinjury. Of particular interest, an additional 22% did so between years 1 and 5. These reports indicate that the prospects for late recovery are better than previously thought and suggest that the time frames for a permanent vegetative state designated by the Multi-Society Task Force should be revisited.
Functional Outcome. Emerging information on longer-term outcome of patients with prolonged disorders of consciousness suggests that long-held beliefs of poor outcome in this patient population are incorrect. Extrapolating from the Multi-Society Task Force report in 1994, a substantial proportion of patients with a vegetative state over 1 month after TBI achieved favorable outcomes. Using the Glasgow Outcome Scale to describe functional outcome at 12 months postinjury, 22% of the surviving patients with TBI remained vegetative, 42% were severely disabled, 25% were moderately disabled and 10% achieved a good recovery level. Therefore, nearly 25% of all patients in a vegetative state over 1 month after TBI and over one-third of survivors achieved substantial independence at home and at least partial independence in the community by 1 year. It should also be recognized that the severe disability category ranges from continuous total dependency to the need for assistance with only one activity; thus, some in this category may have also achieved considerable independence at home, albeit with the need for some assistance. Functional outcome was somewhat worse after non-TBI; however, 9% of survivors achieved a favorable outcome—moderate disability or better—by 1 year. Outcome for children was better at 12 months: one-half were severely disabled while most of the remainder achieved a good recovery. Older adults (> 40 years old) tended toward a worse functional outcome, rarely improving beyond the level of severe disability.
Rate of recovery and functional outcome appear to be substantially more favorable for patients in a minimally conscious state relative to those in a vegetative state evaluated at a similar time postinjury. Giacino and Kalmar investigated functional outcome on DRS across the first year postinjury in patients diagnosed with a vegetative state (n = 55) or a minimally conscious state (n = 49) after TBI (n = 70) or non-TBI (n = 34) on admission to rehabilitation. Although both diagnostic groups presented with similar levels of disability at 1 month postinjury, outcome was significantly more favorable by 12 months in the minimally conscious state group, particularly after TBI. The probability of a more favorable outcome (moderate or no disability) by 1 year was much greater for the minimally conscious state group (38%) than the vegetative state group (2%) and only occurred in those patients with TBI.
Longer-term follow-up studies of patients with prolonged disorders of consciousness have demonstrated considerable proportions achieving favorable outcomes, some with continuing meaningful recovery after 1 or 2 years postinjury. Lammi and colleagues followed 18 patients in a minimally conscious state after TBI for 2 to 5 years after discharge from an inpatient brain injury rehabilitation program in Australia. The authors found that 15% of their sample had partial disability or better at follow-up, while 20% fell in the extremely severe to vegetative category. Similarly, Giacino and Kalmar reported that 23% of their sample had no worse than partial disability at 12 months and 17% classified as extremely severe to vegetative. In both studies, the most common outcome was moderate disability, which occurred in ~ 50% of patients. Lammi et al noted that 50% of their sample had regained independence in activities of daily living.
In a consecutive series of 36 patients admitted to rehabilitation in a vegetative state (n = 11) or a minimally conscious state (n = 25) after TBI (n = 22) or non-TBI (n = 14) and followed for 1 to 4 years postinjury, Katz et al found that 72% emerged from a minimally conscious state and 58% recovered further, clearing from the postinjury confusional state and posttraumatic amnesia (CS/PTA). Patients admitted in a vegetative state took significantly longer than those admitted in a minimally conscious state to emerge from a minimally conscious state (mean 16.4 weeks for a vegetative state; 7.4 weeks for a minimally conscious state) and to emerge from CS/PTA (mean 30.1 weeks for a vegetative state; 11.5 weeks for a minimally conscious state). Patients who failed to clear CS/PTA at latest follow-up were either patients with non-TBI or those with a vegetative state lasting more than 8 weeks. Of those followed over 1 year, nearly half achieved household independence, 22% returned to work or school, 17% at or near preinjury levels.
In the previously discussed study of patients admitted to the TBI Model Systems and followed for up to 5 years, a substantial proportion went on to achieve favorable functional outcomes. Of the 269 patients who regained consciousness, 23% cleared CS/PTA, 21% were capable of living without in-house supervision and 20% were employable based on DRS criteria. Further analysis of a subgroup of these patients (n = 108) stratified to those who recovered consciousness before rehabilitation discharge and those who did not were followed on the Functional Independence Measure (FIM), an observational rating scale that assesses self-care (e.g., toileting), mobility (e.g., walking, transfers), and cognition (e.g., memory, communication, problem solving). Of those with earlier return to consciousness, 56 to 85% achieved independence on one or more functional domains, versus 19 to 36% of those who did not. The earlier recovery group continued to show increasing proportions achieving independence between 2 and 5 years after TBI, whereas the late recovery group did not show substantial return to functional independence after 2 years. This analysis supports the thesis that a substantial proportion of patients with prolonged disorders of consciousness achieve independent functioning through the first 5 years postinjury. Collectively, the aforementioned long-term outcome studies suggest that recovery continues longer than previously believed and indicate that functional outcome is significantly more favorable in a substantial minority of patients with prolonged disorders of consciousness. As many as 20% continue to evidence demonstrable recovery over several years en route to eventually achieving independence at home and in the community. Patients admitted in a minimally conscious state improve faster and have better prospects for recovery of function than those admitted in a vegetative state. Those with traumatic injuries generally have a longer course of recovery and better outcomes overall than those with non-TBI. Table 1 summarizes the TBI Model Systems data on recovery of consciousness and function to illustrate the relatively favorable long-term prospects for improvement in this population.