TRAUMATIC BRAIN INJURY (TBI) is a leading cause of death and permanent disability in the United States. The primary causes of TBIs are falls (28%), motor vehicle accidents (20%), being hit by objects (19%), and assaults (11%).1 TBI can be classified as mild, moderate, or severe, most often by Glasgow Coma Scale score (mild = 13-15; moderate = 9-12; severe = <=8 out of 15). Symptoms of mild TBI (mTBI) include headache, confusion, lightheadedness, dizziness, blurred vision, tinnitus, bad taste in the mouth, fatigue or lethargy, change in sleep patterns, behavioral or mood changes, and trouble with memory, concentration, attention, or thinking.2 In addition to these symptoms, individuals with moderate or severe TBI may also experience headaches that intensify or do not go away, repeated vomiting or nausea, convulsions or seizures, an inability to awaken from sleep, dilation of one or both pupils of the eyes, slurred speech, weakness or numbness in the extremities, loss of coordination, and confusion, restlessness, or agitation.

Sleep disturbance is one of the most common yet least studied of the post-TBI sequelae. Recent research suggests that 30% to 70% of patients experience sleep problems following TBI and that these sleep disturbance often exacerbate other symptoms and impede the rehabilitation process and the ability to return to work.3 Interestingly, it has been shown that it is mTBI that is most frequently associated with sleep disturbance as opposed to severe TBI.4,5 While the importance of sleep problems in the general population of TBI patients has been known for some time, recent events such as the wars in the Middle East have brought renewed attention to this issue. Many veterans have sustained a TBI, and the treatment of sequelae associated with this type of injury has become a major health policy concern.

PURPOSE OF REVIEW

This review addresses the etiology and implications of sleep problems in TBI across 4 general domains of current scientific inquiry and observation: subjective impression of poor sleep; objective changes in sleep-related parameters; alterations in circadian rhythms; and neurophysiologic and/or neuropsychologic abnormalities associated with TBI. Both nonpharmacologic and pharmacologic treatment options for sleep problems in this population are then presented, as well as limitations and areas for future research.

For this review, PubMed and Ovid/MEDLINE databases were searched by using the following key words: "sleep disturbance," "insomnia," "TBI," "brain injury," and "circadian rhythms." Each key word was crossed with the others and the abstracts of manuscripts resulting from this search were reviewed by the first author. The review was limited to those studies that focused specifically on sleep and TBI. Note that many of the studies reviewed included patients with TBI of varying severities, and they typically did not report results separately for each severity type. Intervention studies were included provided that the primary aim of the study was not limited to comparing effectiveness of pharmaceutical agents. Additional studies were sourced from abstracts from the Associated Professional Sleep Societies meetings and published in the journal Sleep over the past decade. Also, only articles published in English language journals over the past decade (since 1997) and those that could be searched electronically were included in the review. In all, 33 articles met these criteria and were selected for this review (Table 1). Articles were later grouped by domain of inquiry for easier synthesis and readability. While it is possible that some studies were missed with this search method, this review provides a broad overview of the current state of knowledge regarding sleep disturbance and its relation to TBI and recovery.

Table 1 Traumatic brain injury and sleep disturbance-summary of reviewed articles

Table 1 Traumatic brain injury and sleep disturbance-summary of reviewed articles (Continued)

Table 1 Traumatic brain injury and sleep disturbance-summary of reviewed articles (Continued)

TBI AND SUBJECTIVELY REPORTED SLEEP DISTURBANCE

One of the most common and well-documented syndromes seen in patients with TBI is insomnia, a disorder inherently defined by patient self-report. The diagnostic criteria for insomnia are a problem of initiating and/or maintaining sleep for at least 1 month accompanied by subjective impairment in daytime functioning.37 Evidence showing that insomnia is a problem for TBI patients has come from several studies. For example, in a questionnaire study of 452 TBI patients, up to 50% of individuals reported insomnia symptoms and 29.4% fulfilled the diagnostic criteria for an insomnia syndrome.6 For those who met criteria for insomnia, sleep problems were a severe, chronic, and untreated condition in almost 60% of cases. Risk factors associated with insomnia were mTBI, higher levels of fatigue, depression, and pain. In another study of 50 post-acute TBI admissions and a comparison group of 50 rehabilitation outpatients, 30% of all patients were found to suffer from insomnia, while an additional 12% experienced a degradation of sleep quality as measured by the Pittsburgh Sleep Quality Index.7 In this sample, sleep initiation was a problem almost twice as often as sleep duration.

In a longitudinal study of admissions to a rehabilitation unit, Clinchot et al8 found that 50% of patients reported difficulty sleeping, 64% reported waking up too early, 25% described sleeping more than usual, and 45% complained of problems falling asleep. Overall, 80% of subjects reporting sleep problems also reported problems with fatigue. Furthermore, in a study by Parcell and colleagues,9 63 patients with TBI recruited after discharge from rehabilitation and 63 age- and sex-matched controls from the general community were compared on sleep variables. Patients were shown to have a significantly higher frequency of reported sleep disturbance following their TBI (80%), and they reported more nighttime awakenings and longer sleep-onset latency. In comparison, sleep problems were only reported by 23% of individuals in the control group. In keeping with other studies, participants with a more mild form of TBI reported sleep changes more frequently.

Insomnia not only represents a problem immediately following injury but also appears to persist for months and years after a TBI. For example, in a review of medical records from 175 patients diagnosed with mTBI taken at 2 time intervals (interval 1 = mean 10.7 days vs interval 2 = mean 6.3 weeks postinjury), Chaput et al13 found that complaints of sleep disturbance were reported by 11.1% and 34.7% of patients. Overall, there was a 2-to 3-fold increase in the prevalence of sleep disturbance from 11 days to 6 weeks postinjury. In a study of sleep complaints in 39 patients with chronic postconcussion syndrome and a control group of 27 patients with orthopedic injuries, the former reported more difficulty initiating and maintaining sleep at night and greater difficulty with sleepiness during the day over a mean 2-year postinjury period.10

Among 19 adolescent patients who were 3 years post-mTBI without any discernible clinical sequelae, complaints of sleep disturbance could be confirmed by both polysomnographic (PSG) and actigraphic (ie, daily activity) monitoring.11 In addition, in a study of 22 hospitalized patients with TBI of recent onset (median 3.5 months postinjury) and 77 discharged patients who had sustained brain injury about 2 to 3 years earlier (median 29.5 months), Cohen et al12 found high rates of sleep complaints in both groups (72.7% and 51.9%, respectively). Difficulty initiating and maintaining sleep were the most common complaints among hospitalized patients (81.2%), whereas excessive daytime somnolence was most common in discharged patients (72.5%). Notably, those discharged patients who reported sleep complaints also reported greater neurobehavioral impairments and poorer occupational outcomes.

Among the typical consequences of poor nighttime sleep in any population are complaints of daytime fatigue or excessive sleepiness. Not surprisingly then, there is also a relationship between sleep disturbance and these daytime symptoms in TBI patients. Fatigue may be present in as many a 43% to 73% of patients and is one of the primary symptoms in 7% of patients regardless of injury severity, time since injury, gender, or medications.14,15 Lundin et al16 found that poor memory, sleep disturbance, and fatigue were the most commonly reported symptoms following TBI. While these complaints declined postinjury, correlations between symptoms and disability were still evident at 3 months postinjury. Furthermore, patients with early high symptom load were found to be at risk for developing persisting complaints. In a study by Baumann and colleagues,17 33 of 53 patients (62%) were found to have a sleep disorder following TBI, of whom 29 (55%) reported increased fatigue or excessive daytime sleepiness compared with pre-TBI. In this study, Epworth Sleepiness Scale scores (a measure of trait sleepiness) were abnormally elevated (>10) in 14 patients (26%), suggesting that many patients exhibited high levels of daytime sleepiness (especially those with mTBI as compared with those with moderate/severe TBI).

Overall, subjectively reported sleep disturbance is one of the most commonly described problems in TBI patients. The literature reviewed in this section is particularly convincing because numerous studies seem to be providing replicable findings of insomnia in TBI patients. Most of the studies surveyed (especially the study of Ouellet and Morin)24 had reasonably large sample sizes and utilized a variety of measures to assess sleep disturbance. In addition, daytime fatigue and sleepiness were also common findings in these studies. Interestingly, it is the research on subjectively defined sleep problems that has brought attention to the observation that mTBI is more clearly associated with insomnia than severe forms of TBI. While the exact cause of subjectively reported sleep difficulties and daytime sleepiness is not known, these problems obviously affect patients both cognitively and psychologically as they recover from their injuries and impede attempts to return to normal activities.

TBI AND OBJECTIVELY MEASURED SLEEP DISTURBANCE

Despite the wealth of information supporting the presence of subjective sleep complaints in TBI patients, evidence for measurable objective deficits in sleep has been harder to identify. Some evidence of objective changes in sleep quantity and quality has been found by Kaufman and colleagues11 in a study of 19 adolescents (3 years post-mTBI) who complained of sleep disturbance and a group of control participants. Questionnaire results revealed that patients subjectively reported more severe sleep complaints, a result that was corroborated by PSG measurements. In fact, PSG recordings revealed that in comparison with controls, mTBI was associated with lower sleep efficiency, more nocturnal wake time, and more awakenings lasting longer than 3 minutes. In a study by Schreiber and colleagues,18 26 mTBI patients with normal brain computed tomographic and negative encephalographic (EEG) studies were compared with a matched group of healthy controls. Patients in this study also showed altered sleep architecture, with significantly higher percentage of nonrapid eye movement (NREM) stage 2 than did controls and significantly lower percentage of rapid eye movement (REM) sleep. In addition, Gosselin et al19 found evidence of objective changes in the EEG during wakefulness in athletes with concussion relative to control athletes. In particular, athletes with concussion were shown to have significantly more delta activity and less alpha activity during wakefulness. This pattern has been seen in sleep-deprived and sleep-disordered populations and suggests that abnormal waking EEG may play a significant and/or unique role in the symptoms associated with TBI, especially daytime dysfunctions and sleepiness experienced in this population.

Of note, some patients with sleep disturbance after TBI may experience sleep problems other than insomnia. Lai et al20 found an increased risk of sleep disorders, including sleep apnea. In addition, several patients in this study also had a prior history of sleep disorders, which the authors suggest might have contributed to the TBI. Castriotta et al21 and Verma et al22 have similarly reported high numbers of sleep disorders such as obstructive sleep apnea, periodic limb movements, and narcolepsy in patients with TBI. Such findings have led some to propose that preexisting sleep disorders and subjective daytime sleepiness may be more common in populations with brain injury and may actually represent a risk factor for accidents or behaviors that predispose to head injury.23

Contrary to these findings, several studies have failed to find clear evidence of objective sleep impairments in patients with mTBI. For example, Ouellet and Morin24 found that TBI patients with insomnia, relative to healthy controls, showed a tendency to overestimate the subjective impression of their sleep disturbance as compared with PSG measurements. Despite reports of significant sleep disturbance on subjective measures, patients showed no differences from controls in percentage of stage 2, slow-wave (stages 3 and 4), or REM sleep on PSG measurements. Patients did, however, have a higher proportion of stage 1 sleep, more awakenings lasting longer than 5 minutes, and shorter REM sleep latencies. Such a pattern of mild sleep disturbance that is overestimated on subjective measures is consistent with the overall insomnia literature, suggesting that the experience of insomnia may be similar in persons with and without TBI. In a study by Aurora and colleagues,25 5 men and 4 women (aged 24-53 years) with a median time from injury of 7 years were studied with PSG measurements. Results indicated that subjects had normal sleep efficiencies (median of 90.7%); the median amounts of stage 1, stage 2, slow-wave, and REM sleep were 1.9%, 32.5%, 43.8%, and 18.0%, respectively. The authors concluded that while TBI patients may suffer from subjective complaints of fatigue and poor nocturnal sleep, they nevertheless manifest relatively normal sleep architecture.

Several other investigations also seem to support this observation of a disparity between subjective and objective measures of sleep disturbance following TBI. For example, Korinthenberg and colleagues26 studied 98 children (aged 3-13 years) within 24 hours after a minor head injury and 4 to 6 weeks later. At follow-up, 23 of 98 still exhibited posttraumatic complaints with headache, fatigue, sleep disturbance, anxiety, and affect instability. These posttraumatic symptoms, however, did not correlate with somatic, neurologic, or EEG findings observed immediately after the injury or at the follow-up investigation. Chung et al27 found that subjective reports of daytime sleepiness and fatigue in patients with head injury were not well substantiated when compared with overnight PSG sleep measures and physiologic tests of daytime sleepiness.

Despite containing nearly an equal number of manuscripts as the section on subjective sleep disturbance, the evidence for objective sleep problems in mTBI is the most equivocal of any of the outcome measures reviewed here. Inconsistencies in the literature may be due to multiple factors such as (1) small sample sizes in most studies (with a few notable exceptions), (2) more variability in the age ranges across studies (including a larger number of children and adolescents), and (3) use of samples with mixed severity of injury. On the other hand, the variability in TBI severity and the time since injury are about the same in this set of articles as is in the reports on subjective sleep disturbance. However, subjective and objective discrepancies in sleep disturbance, such as those observed in TBI patients, are not at all inconsistent with the general literature on insomnia. Nonetheless, the largest studies in this section suggest that TBI patients do experience objective sleep disturbance and/or sleep disorders postinjury.21,22,26 The other intriguing hypothesis that warrants more longitudinal research is the notion that preexisting sleep disturbance may be associated with increased risk of TBI.23

TBI AND CIRCADIAN CHANGES

Recent research has provided evidence that, rather than insomnia per se, sleep disturbance following TBI may be associated with alterations in the timing and rhythm of sleep. Such alterations can result in a mismatch between an individual's biological sleep-wake schedule and his or her desired 24-hour environmental and social schedule.38,39 Of the circadian dysrhythmias, delayed sleep phase syndrome is the most common and is often misinterpreted as insomnia by patients and clinicians alike. In fact, the estimated prevalence of circadian disorders among patients who initially complain of insomnia is 7% to 10%.38,39 Patients with circadian rhythm sleep disorders (CRSDs) often exhibit altered rhythms of melatonin secretion and body temperature.40,41 Consistent with these notions, recent case studies of CRSDs following TBI have reported evidence of delayed sleep phase,28 delayed sleep phase with delayed melatonin and body temperature rhythms,29 and a non-24-hour sleep-wake pattern with abnormal melatonin rhythm.42 Recent work by Ayalon and colleagues30 found that mTBI might actually contribute to the emergence of CRSDs. In this latter study, 2 types of disturbances were observed: delayed sleep phase syndrome and irregular sleep-wake pattern; these types differed in subjective questionnaire scores and had distinct profiles of melatonin and temperature circadian rhythms. The authors also reported that 15 of 42 patients (36%) with complaints of insomnia following mTBI were more properly diagnosed with CRSDs. The frequency of CRSDs in this sample is considerably higher than the prevalence of these disorders among the general population of patients attending sleep clinics who complain of insomnia.38,39

Not all research, however, supports the idea that TBI leads to circadian disturbances. In a study of 10 patients in a post-acute TBI group compared with matched controls, Steele and colleagues31 found that the TBI and control groups reported similar habitual sleep times as measured with a Morningness-Eveningness questionnaire. Furthermore, the timing of melatonin onset did not differ between the groups. While the small sample size may have contributed to the lack of group differences, the preliminary conclusion of the authors is that this study failed to provide conclusive objective evidence of a shift in circadian timing of sleep following acute TBI.

In summary, a growing body of evidence suggests that for some patients, circadian disturbances may play a role in sleep problems subsequent to TBI. Unfortunately, the literature in this area is sparse, and the 2 primary investigations reviewed here (studies of Ayalon et al and Steele et al)30,31 came to opposite conclusions as to the role of circadian dysrhythmia in TBI patients. It is obvious that future studies are needed on this connection and its limits.

TBI AND NEUROPHYSIOLOGIC/NEUROPSYCHOLOGIC DISTURBANCES

Recent research is beginning to demonstrate that TBI may be implicated in neurophysiologic changes that affect the regulation of sleep and wakefulness. For example, Baumann and colleagues32 studied 44 consecutive patients with acute TBI and controls and found hypocretin-1 levels to be abnormally low in 95% of patients with moderate to severe TBI and in 97% of patients with posttraumatic brain changes as indicated by computed tomographic scan. In a follow-up study, Baumann and colleagues33 observed low cerebral spinal fluid hypocretin-1 levels in 4 of 21 patients 6 months following a TBI (mixed severities) as compared with 25 of 27 patients in the first days after TBI. Furthermore, significantly lower levels of hypocretin-1 were associated with posttraumatic excessive daytime sleepiness. These data suggest that hypocretin-1 deficiency after TBI may reflect hypothalamic damage in patients with acute TBI. Such deficiencies would be expected to contribute to sleep disturbance via hypocretin's role in promoting wakefulness (eg, the study of Mignot et al).43 More specifically, abnormally low levels of hypocretin would be expected to contribute to daytime sleepiness and fatigue, 2 common complaints in patients with mTBI.

Sleep-endocrine alterations have also been shown to be present several months after severe TBI and exhibit a pattern of sleep EEG parameters and nocturnal hormone secretions similar to that seen in patients with remitted depression.34 Frieboes et al34 propose that hypothalamic-pituitary-adrenal overdrive and long-term modulation of hypothalamic and pituitary receptors may lead to either permanent sleep-endocrine alterations (a neurobiologic "scar") or perhaps hypothalamic-pituitary damage due to diffuse thinning out of neurons (eg, growth hormone secreting cells). Similar research has shown that patients with TBI are more likely than healthy, age-matched peers to report problems with their metabolic/endocrine and neurologic systems.35

Current research also indicates that deficits in cognitive performance may be related to the degree of sleep disturbance resulting from TBI. In a study of 87 patients with mild to severe TBI admitted to a comprehensive outpatient neurorehabilitation program, hierarchical regression analysis revealed that poor performance on selected measures of cognitive functioning was significantly related to sleep disturbance, accounting for 14% of variance beyond that accounted for by injury severity and gender.4 These results suggest that sleep disturbance among patients with mTBI may be associated with impaired neuropsychologic functioning. Sleep disturbance has also been shown to correlate with posttraumatic amnesia. In a study of 9 patients in a rehabilitation unit, sleep efficiency was positively correlated with ratings of orientation and significantly predicted clearance of posttraumatic amnesia.36

Overall, it appears that TBI can induce measurable neurophysiologic changes that may lead to objective alterations in sleep-wake regulation. In particular, the works of Baumann et al32 and Friebos et al34 provide convincing evidence that TBI results in endocrine changes that might explain sleep disturbance in this population. In addition, the literature reviewed here suggests that sleep changes may contribute to cognitive performance deficits in persons with TBI. While only a few studies indicated a link among TBI, sleep problems, and neuropsychologic performance, such results should not be surprising, given the established role that sleep plays in cognition. One issue in need of further investigation is the amount of variance in neuropsychologic impairment that can be directly and uniquely attributed to sleep disturbance, independent of other factors.

TREATMENT OF SLEEP PROBLEMS ASSOCIATED WITH TBI

As researchers debate the complex issues surrounding the underlying etiology and course of symptoms resulting from TBI, clinicians must focus on how best to treat sleep disturbance in this population. Overall, the TBI-related sleep disturbance most strongly supported in the literature are consistent with insomnia and delayed sleep phase syndrome. Fortunately, there are many proven psychologic and pharmacologic interventions that can help patients with these sleep disturbance. In fact, several cognitive-behavioral techniques (CBTs) have been effective in improving sleep in patients with insomnia and thus hold particular promise for persons with TBI. Unfortunately, we found only a single report in which a CBT for insomnia was applied with individuals with TBI (ie, the work of Ouellet and Morin).44 In this case study, a patient treated with a CBT had reductions in sleep-onset latencies from 47 to 18 minutes, nocturnal awakenings dropped from 85 to 28 minutes, and sleep efficiencies increased from 58% to 83%. These results were well maintained at 1- and 3-month follow-up assessment. Furthermore, there are several therapies that have been effective in the treatment of CRSDs and which may have applicability in this population. For example, bright light therapy will aid in the consolidation of sleep in patients with CRSDs.44 Given that alterations in melatonin functioning also lead to sleep-phase alterations seen in TBI patients, treatment with exogenous melatonin could be useful in adjusting sleep-phase problems seen in this population.

Despite the effectiveness of CBTs for the treatment of sleep disturbance, prescription and over-the-counter medication remain the most common and accessible forms of treatment for most individuals recovering from TBI. In fact, hypnotic medications may be prescribed in as many as 20% of TBI patients.45 While medications have been effective in treating sleep disturbance in healthy individuals, such treatment effects have not been studied in TBI patients. Use of hypnotic medications may also be associated with daytime drowsiness, dizziness or lightheadedness, and cognitive and psychomotor impairments, all of which are likely to be particularly detrimental for this population.

LIMITATIONS AND FUTURE DIRECTIONS

TBI is a common and economically costly healthcare issue in the United States today, with sequelae that are numerous and include both medical and psychiatric symptoms. It is associated with frequent complaints of sleep disturbance that likely result from and contribute to multiple interrelated factors associated with the injury, all of which complicate recovery and resolution of symptoms. While current research is beginning to elucidate the nature of the relationship between TBI and sleep disturbance, the literature is hampered by many methodologic concerns. As seen in this review, studies over the last decade vary considerably across important factors such as sample size, age, injury severity, types of measures/assessments performed, and the length of time between injury and evaluation. In addition to variability in study design, reporting of important factors such as the number of prior TBIs, patient treatment regimen, medication use, past/current medical and/or psychiatric status, etc varied considerably among articles (and often were not reported at all), making it difficult to evaluate the role of these variables in sleep-related outcomes and complicating comparisons between studies.

There are, of course, several other limitations to the literature. For example, very little attention has been paid to symptoms associated with a specific type or location of injury, making logical cross-comparisons of studies almost impossible. Second, studies have not adequately investigated the effects of TBI on sleep on the basis of time since injury (immediate vs long term). Third, the relative role of physiologic and/or psychologic factors in the emergence of sleep disorders (ie, due to physical injury in sleep/arousal brain regions vs trauma/stress) has not been fully explored. Fourth, whether impaired sleep is the result of or the cause of other TBI-related symptoms and whether impaired sleep may even increase likelihood of injury is yet to be determined. Fifth, TBIs typically result in multiple sequelae, such as pain, mood problems, posttraumatic stress disorder, and sleep disturbance. Hence, teasing apart comorbid symptoms and studying their contributory effects to illness and recovery are difficult processes at best. Finally, for obvious reasons, premorbid sleep patterns in individuals who suffer a TBI are rarely assessed. Clearly, knowledge of such status might have implications for the understanding of post-TBI functioning.

Another important limitation is the inability of research to explain why mTBI seems to be more commonly associated with sleep disturbance than more severe forms of TBI. While this fact may seem counterintuitive, it does appear that mTBI is correlated with increased likelihood of insomnia or delayed sleep phase syndrome. While no clear explanation for this link has been established, one could speculate as to the potential reasons for its occurrence. First, it may simply be a fact that more severe brain injuries result in more complex issues during the recovery process of patients. In such cases, rehabilitation is focused on matters such as reestablishing the ability to conduct activities of daily living and reduction of long-term disability. Thus, sleep disturbance is less likely to become a focus of treatment and/or get reported. Second, differences in the nature of injuries may play a role in the development of sleep disturbance. Perhaps, more diffuse injuries (eg, axonal shearing) that occur in mTBI lead to impairment in global functioning (eg, arousal) more so than acute or focal traumas. Third, severe injuries may produce more permanent changes in brain function, as opposed to mild injuries, and this difference may alter in some meaningful way the recovery process, causing sleep problems to occur more often in one group than another. Finally, differences in the way brain injuries are treated may lead to differences in self-reported postinjury symptoms. Recent research by Chaput et al13 has found that individuals with mTBI were more likely to report mood and sleep problems 6 weeks after injury than in the time immediately following the injury. The author (personal communication) concluded that sleep problems manifested at a later time because patients with mTBI are often rapidly discharged from care (as opposed to those with more severe injuries). In such cases, the appearance of sleep problems later in the course of recovery might be due to patients running out of medication and not receiving adequate follow-up services.

In conclusion, it is obvious that TBI is associated with sleep disturbance. However, it is also clear that much more research needs to be done to understand the extent of this relationship. More important, research needs to focus on uncovering the specific types, causes, and severity of TBI that most often lead to sleep problems, as well as the most appropriate treatment strategies for the resultant sleep-wake abnormalities. It is hoped that recent public attention regarding the significance of TBI will spur support of research that is essential both to improving conceptualizations of this complex issue and to improving therapeutic strategies that could benefit those who suffer from this debilitating condition.

REFERENCES