Natural history of recovery from brain injury after prolonged disorders of consciousness: outcome of patients admitted to inpatient rehabilitation with 1–4 year follow-up

Abstract

The natural history of recovery from brain injury typically consists of a period of impaired consciousness, a subsequent period of confusion and amnesia, followed by a period of post-confusional recovery of function. Patients with more severe injuries may have more prolonged episodes of unconsciousness or minimal consciousness and may not fully evolve through this continuum of recovery. There is limited information on the course of recovery and long-term outcome of patients with prolonged unconsciousness, particularly those with extended periods in the minimally conscious state. Further, patients with impaired consciousness are frequently denied access to hospital-based rehabilitation services because of uncertain prognosis and a perceived lack of benefit from rehabilitative interventions.

Methods: A consecutive series of 36 patients with traumatic (TBI) and non-traumatic brain injury (nonTBI) in a vegetative state (VS) or minimally conscious state (MCS) on admission to a specialized, slow-to-recover brain injury program in an acute rehabilitation hospital was retrospectively reviewed to evaluate course of recovery during rehabilitation hospitalization and in follow-up, 1–4 years post-injury. Independent variables included: time to resolution of VS, MCS and confusional state/posttraumatic amnesia (CS/PTA), based on Aspen criteria, Coma Recovery Scale-Revised (CRS-R) and Galveston Orientation and Amnesia Test (GOAT) scores. Outcome measures (calculated separately for TBI, nonTBI, VS, or MCS on admission subgroups) included: proportion of patients who recover and recovery time to MCS, CS/PTA stages, household independence, and return to school or work, as well as Disability Rating Scale (DRS) scores at 1, 2, 3, and 4 years post-injury.

Results: The majority emerged from MCS (72%) and CS/PTA (58%) by latest follow-up. It took significantly longer for patients admitted in VS (means: MCS, 16.43 weeks; CS/PTA, 30.1 weeks) than MCS (means: MCS, 7.36 weeks; CS/PTA, 11.5 weeks) to reach both milestones. Almost all who failed to clear CS/PTA by latest follow-up were patients with nonTBI or TBI with VS lasting over 8 weeks. Duration of MCS was a strong predictor of duration CS/PTA after emergence from MCS, accounting for 57% of the variance. Nearly half the patients followed at least 1 year achieved recovery to, at least, daytime independence at home and 22% returned to work or school, 17% at or near pre-injury levels. Discharge FIM score or duration of MCS, along with age, were best predictors of DRS in outcome models. DRS scores continued to improve after 2 and 3 years post-injury.

Conclusions: Patients in VS whose transition to MCS occurred within 8 weeks of onset are likely to continue recovering to higher levels of functioning, a substantial proportion to household independence, and productive pursuits. Patients with TBI are more likely to progress than patients with nonTBI, though significant improvement in the nonTBI group is still possible. Active, higher intensity, rehabilitation should be strongly considered for patients with severely impaired consciousness after brain injury, especially for patients with TBI who have signs of progression to the MCS.

Introduction

Patients with traumatic brain injury (TBI) typically progress in recovery from a period of impaired consciousness to a posttraumatic confusional state with amnesia, to a period of post-confusional improvement of attention, memory, and executive capacities (Povlishock and Katz, 2005). This pattern is observed across a broad range of TBI severity. Patients with non-traumatic brain injury (nonTBI) may have a similar pattern of recovery, though prognosis is usually worse within the same range of severity (Multi-Society Task Force on PVS, 1994). The natural history of recovery of TBI has been further characterized by a series of clinically defined conditions or stages of recovery that have been labeled according to various schemas, including, the one most commonly used for TBI, the Rancho Los Amigos (RLA) Scale of cognitive recovery (Table 1) (Hagen et al., 1972). The first three levels on the RLA Scale describe unconsciousness and emerging consciousness. The terms coma, vegetative state (VS), and minimally conscious state (MCS) largely correspond to these first three levels on the RLA Scale. The posttraumatic confusional state and posttraumatic amnesia (CS/PTA) are included in the next three levels, Rancho 4–6, and the post-confusional period corresponds to levels 7 and 8. Another schema describes these stages using some of the more familiar neurologic nomenclature (Table 2) (Alexander, 1982; Katz, 1992; Povlishock and Katz, 2005). The transition from coma to VS, which occurs within 2–3 weeks in the vast majority of survivors, is marked by spontaneous eye opening. The transition from VS to MCS, is defined by the first signs of minimal, inconsistent, but reproducible behavioral evidence of self or environmental awareness, as defined by the Aspen workgroup criteria (Giacino et al., 2002). The transition from the MCS to the next stage, labeled CS/PTA is marked by the Aspen workgroup criteria of accurate yes/no communication or object use. Sometimes object use and functional communication return simultaneously, and sometimes one or the other criterion returns first (Giacino and Kalmar, 2005; Taylor et al., 2007). Transition to the next stage, post-confusional/emerging independence is marked by clearing of PTA that can be designated using standardized measures such as the Galveston Orientation and Amnesia Test (GOAT) (Levin, 1979). The transition to the last stage, community reentry/social competence is defined in this schema by achievement of daytime independence at home, the ability to be left alone for an 8-h period.

Patients with TBI may progress through these stages of recovery at different rates, largely depending on injury severity; not all stages will be clinically recognized in every patient. Patients with more severe injuries may have more prolonged episodes of coma, VS, MCS or CS/PTA and may not fully evolve through this continuum of recovery. There is some information about the probability of recovery beyond the VS if it lasts a month or more; but there is more limited information on the probability of recovery beyond a MCS that extends a month or more.

For patients in a prolonged VS, the evolution and probability of recovery for survivors has been described in a meta-analysis of persistent vegetative state (defined by the Multi-Society Task Force on PVS as those fully unconscious for a month or more) (Multi-Society Task Force on PVS, 1994). Prognosis for recovery was substantially better for victims of TBI than those with nonTBI. The report described functional outcome using the Glasgow Outcome Scale (Jennett and Bond, 1975). Of adults with TBI who were unconscious at least 1 month, 33% recovered consciousness by 3 months post-injury, 46% by 6 month, and 52% by 1 year. If patients with TBI were still unconsciousness at 3 months, 35% regained consciousness by 1 year; if still unconscious for 6 months, 16% regained consciousness by 1 year. Of those adults with nonTBI who were unconscious for 1 month, only 11% recovered consciousness by 3 months, and 15% by 6 months. No person with nonTBI regained consciousness after 6 months post-injury. Functional outcomes at 12 months for patients with TBI who regained consciousness were as follows: more than 1/2 were severely disabled, nearly 1/3 were moderately disabled, and about 13% reached a good recovery level. Functional outcome was worse after nonTBI. Nearly 3/4 of those who regained consciousness were severely disabled at 12 months, though 1/5 were moderately disabled, and 1/15 achieved good recovery (Multi-Society Task Force on PVS, 1994).

There is less information available for patients in a prolonged MCS. One study (Giacino and Kalmar, 1997) compared outcomes of patients in VS versus MCS admitted to a rehabilitation facility. In this study 55 patients in a VS were compared to 49 patients in a MCS when initially evaluated a mean of 9.6 weeks post-injury. Causes of injury were TBI (n=70) and nonTBI (n=34) (mostly anoxic brain injury and stroke). Using the Disability Rating Scale (DRS) (Rappaport et al., 1982) as the outcome measure at 1, 3, 6, and 12 months post-injury, they reported that the probability for the most favorable outcomes (moderate or no disability) by 1 year was much greater for the MCS group (38%) than the VS group (2%) and only occurred in those patients with TBI. Of the MCS group, 43% remained severely disabled or worse (1/10 of the nonTBI, MCS group was vegetative and 2/10 died) at 12 months.

Another study of 18 patients with TBI admitted to rehabilitation in a MCS of at least 27 days duration (median 56 days), followed them for 2–5 years (Lammi et al., 2005). Two patients persisted in a MCS at follow-up, 4 or more years after injury. On the DRS, 1 was mildly disabled (DRS=1), 2 partially disabled (DRS=2), 11 moderate to moderate/severe (DRS 4–11), and 4 extremely severe or vegetative (DRS 25–30). There was no significant correlation between the duration of MCS and outcome on the DRS or the FIM but there was a correlation with level of cognitive impairment on the Dementia Rating Scale (Schmidt et al., 2005). Of 14 working full-time prior to injury, four returned to part-time work at latest follow-up. Overall, the authors concluded that outcome after prolonged MCS following TBI was heterogeneous and difficult to predict.

Although the previous study fell short of developing a prognostic model, a study of 124 patients admitted among several rehabilitation facilities in either VS or MCS, at least 4 weeks post-injury, did demonstrate significant outcome prediction models for outcome over a shorter interval (Whyte et al., 2005). Level of initial disability (on the DRS), rate of early DRS change and time from injury to initial assessment were the best predictors of the level of disability at 4 months post-injury and time to begin following commands, for those who were not following commands at the initial assessment.

Most of these outcome studies were performed with patients admitted to hospital-level rehabilitation facilities with specialized programs for patients with prolonged impairments of consciousness. It remains unclear what proportion of surviving patients with extended periods of impaired consciousness are treated in such specialized facilities, as opposed to long-term care nursing facilities, home care or rehabilitation facilities without specific expertise in treating this population. In the United States, public and private payers for health services have traditionally considered persons with prolonged impairments of consciousness inappropriate candidates for active rehabilitation assessment and treatment and they are often denied admission to hospital-level rehabilitation facilities. This is in part because traditional admission criteria require active engagement of the individual for a minimum of 3 h/day. Some consider it costly and wasteful to admit patients, who cannot actively interact with therapists, to hospital-level rehabilitation programs. Further, when considering patients who are unconscious or minimally conscious for several weeks after injury, decisions for care are often based on the conclusion that prognosis is uncertain and that the prospect for meaningful recovery is highly unlikely. As a result, once medical problems are stabilized in acute or chronic hospital-level treatment, many patients in VS or MCS remain in long-term nursing facilities, without specialized assessment and rehabilitative care. More cost-effective, intermediate levels of rehabilitative care facilities (subacute rehabilitation, transitional medical rehabilitation) have been proposed (Walker et al., 1996) but few such facilities exist for this population. Once admitted to skilled nursing facilities, patients are unlikely to be transferred to hospital-level rehabilitation facilities. In a large sample of patients with very severe TBI at a low level of functioning, only 3% of those admitted to hospital-level rehabilitation were in a long-term care facility between the acute hospital and hospital-level rehabilitation (Whitlock and Hamilton, 1995). There is very little information on what proportion of patients with severe disorders of consciousness are admitted to hospital-level rehabilitation facility versus a skilled nursing facility after discharge from the acute hospital. There are a variety of clinical and non-clinical factors that influence these admission decisions (Buntin, 2007; Ottenbacher and Graham, 2007). Health insurance was a factor in one study that found that patients with moderate-to-severe TBI were more likely to be admitted to skilled nursing facilities if they had Medicaid or an HMO, as opposed to a fee-for-service plan (Chan et al., 2001).

It remains uncertain what effect treatment at different levels of care or in dedicated programs for patient with impaired consciousness may have on outcome in this population of patients. Although they did not separately evaluate patients with severe impairments of consciousness, a survey of 1059 patients with TBI in Colorado, tracked for their pathway of rehabilitative treatment (inpatient rehabilitation vs. long-term care vs. community-based care) and outcome, found that those who were treated in long-term care facilities had worse outcome at 1 year post-injury at any severity of injury (Mellick et al., 2003). Nevertheless, it is difficult to draw conclusions from such studies with regard to whether placement in a particular level of care is the cause or effect of the level of disability. It is still a realistic concern that a lack of rehabilitation services or inability of less experienced clinicians to recognize subtle or inconsistent manifestations of emerging consciousness may significantly and permanently reduce the potential for recovery of patients with severely impaired consciousness. Indeed, the chance of misdiagnosing patients who are in a MCS as being in a VS is as high as 40% in non-specialized centers (Andrews et al., 1996; Childs et al., 1993). If small signs of emerging consciousness are missed, patients are much less likely to receive active rehabilitation services aimed at promoting further recovery by engaging patients with limited capacities. As more time passes, the chance that payers would approve active rehabilitation services in specialized programs becomes more remote.

Included in this report is an observational study of a cohort of patients who were evaluated in a specialized, inpatient brain injury rehabilitation program, for patients with prolonged disorders of consciousness. Almost all continued to receive some rehabilitation services in other institutions, at home or in outpatient facilities after discharge from the program. The purpose of this study is to better characterize the natural history of recovery and outcome from prolonged disorders of consciousness after brain injury by examining the path of recovery through the stages described above and assessing predictors of long-term functional outcome. Patients with TBI or nonTBI were either in a VS or MCS at the time of admission to this inpatient rehabilitation program. Although this study is not designed to assess the individual contributions of such specialized rehabilitation programs to recovery, it aims to further enlighten awareness of the range of possible outcomes for patients with prolonged disorders of consciousness who are provided active rehabilitation services.