Prevalence of sleep disturbances, disorders, and problems following traumatic brain injury: A meta-analysis

Abstract

Background

Sleep is often disrupted following a traumatic brain injury (TBI), which may compromise recovery and quality of life. Prevalence rates vary widely, reflecting differences in the criteria and measures that are used to assess sleep, as well as sample differences. This meta-analysis examined the prevalence of general and specific, and formally and informally diagnosed, sleep disturbances following TBI in order to establish the nature and extent of these sequelae and their potential impact on recovery.

Methods

Data from 21 studies, which assessed (1) sleep disturbances, regardless of type or severity, (2) diagnosed sleep disorders, and (3) specific sleep problems following TBI, were analyzed and compared to data for the general population.

Results

Overall, 50% of people suffered from some form of sleep disturbance after a TBI and 25–29% had a diagnosed sleep disorder (insomnia, hypersomnia, apnea) – rates that are much higher than those seen in the general population. They were also two to four times more likely to experience problems with sleep maintenance and efficiency, nightmares, excessive sleepiness, early awakenings, and sleep walking.

Conclusion

Sleep disturbances are very common after TBI and have the potential to seriously undermine patient rehabilitation, recovery, and outcomes; making it important to routinely screen for such problems in order to assess both treatment needs and their potential impact on recovery and outcome.

Introduction

Sleep is often disrupted following a traumatic brain injury (TBI) [1], [2], [3] and may cause or intensify a variety of co-morbidities, such as depression, anxiety, irritability, fatigue, cognitive deficits, pain, and functional impairments [3], [4]. In addition, disrupted sleep may compromise a person’s recovery and return to pre-injury activities, and reduce a person’s quality of life [2], [3], [4], [5]. Disrupted sleep may arise from trauma-induced physical and biochemical changes [3] or co-morbid conditions, such as depression [6], and may be impacted by pre-injury sleep problems [7]. Indeed, disrupted sleep is also very common and often overlooked in the general community, and is associated with a wide range of health, social, and economic costs [8]. However, the frequency and type of sleep disturbances experienced following a TBI, and the extent to which they differ from those of the general community, have yet to be clearly established.

Estimates of the prevalence of post-TBI sleep disturbances range between 30% and 84% [9], [10], [11], [12]. The large variability in these estimates limits their clinical utility and is likely to reflect differences in the definitions (i.e., broad vs specific), criteria (e.g., formal diagnosis or not), types of measures (e.g., subjective vs objective), and sources of information (e.g., self-report vs observation) that are used to identify sleep disturbances, combined with differences in the samples that are studied (e.g., non-selected vs symptomatic samples; mild vs severe TBI). A more detailed examination of post-TBI sleep disturbances, which considers these sources of variability, is therefore needed in order to develop clinically informative estimates of the prevalence and type of sleep disturbances that occur after TBI. Moreover, these prevalence rates need to be considered in the context of community base-rates for sleep disturbances.

Research examining changes to sleep following TBI has conceptualized these problems in a number of different ways. Some studies take a broad approach, using either formal or informal criteria; reporting the proportion of people who have any form of disturbance to their sleep, regardless of its type or severity [1], [9], [10], [13], [14]. Other studies report the number of people who have sleep disorders (e.g., insomnia, obstructive sleep apnea) that meet recognized diagnostic criteria, such as the International Classification of Sleep Disorders (ICSD-I, ICSD-R, ICSD-II, American Academy of Sleep [AAS]) [15], [16], [17] or the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV, American Psychiatric Association [APA]) [18]. Finally, there are studies that report the frequency of specific sleep problems, such as problems with initiating or staying asleep, as well as symptoms that may be suggestive of sleep disorders but were not identified using accepted diagnostic criteria (e.g., insomnia, snoring, or excessive daytime sleepiness) [2], [19], [20], [21], [22].

Not only do the definitions and criteria for identifying sleep disturbances following TBI differ between studies, so too do the measures that are used to assess sleep (subjective vs objective) and the methods by which the data are obtained (self-report, clinical interview, and observation). Indeed, sleep data have been collected using: self-report questionnaires [20], [23], such as the Pittsburgh Sleep Quality Index [24] and Epworth Sleepiness Scale [25]; structured diagnostic [26], [27] or clinical interviews [28], [29]; observational measures obtained while a person is sleeping [14]; or objective measures [9], [30], which include recordings of sleep-wake cycles obtained in sleep laboratories (polysomnography).

These different methodologies are likely to have a significant impact on estimates of sleep disturbances following TBI. For example, many people who complain of insomnia are found to have normal sleep patterns when objectively assessed [3], possibly due to the subjective nature of these problems [2], [3] or the impact of compensation claims on self-reported problems [9]. In contrast, people who are formally diagnosed with obstructive sleep apnea often fail to recognize the problem, instead tending to describe poor daytime vigilance [3]. Similarly, large differences have been found between the frequency of problems identified via self-report and those found via clinical interview [2], [29]. Thus, the various methods that are used to collect data can yield different estimates of the frequency of sleep disturbances after TBI and need to be considered separately. Moreover, the fact that a person believes that their sleep is disrupted in the absence of evidence that it is does not necessarily negate the impact of the problem. Rather, different treatments may be required in order to optimize a person’s quality of life and well-being.

Finally, there are variations in the samples that are investigated, particularly in relation to where participants are recruited. Participants with a TBI who are seeking a diagnosis or treatment for sleep problems are likely to have more problems than those who are recruited on a prospective/non-selected basis. Similarly, people who are referred for formal sleep laboratory investigations are likely to be a biased sample, as only those people with more severe problems tend to be referred for specialist investigations of this type. In addition, between-study differences in injury severity and the post-injury time interval may also impact on the nature and frequency of sleep disturbances that are reported.

The current study was designed to examine the frequency and type of sleep problems that occur following TBI, taking into consideration the breadth of problems that were reported (presence of any sleep disturbance vs diagnosed sleep disorders vs specific sleep problems) and the method by which data was obtained (self-report, clinical interview, observation, laboratory tests). Where possible, sample characteristics (symptomatic vs non-selected TBI sample) and the severity of injury (mild, moderate, severe) were also considered.