Sleep Disturbances, TBI and PTSD: Implications for Treatment and Recovery

Abstract

Post-Traumatic Stress Disorder (PTSD), traumatic brain injury (TBI), and sleep problems significantly affect recovery and functional status in military personnel and Veterans returning from combat. Despite recent attention, sleep is understudied in the Veteran population. Few treatments and rehabilitation protocols target sleep, although poor sleep remains at clinical levels and continues to adversely impact functioning even after the resolution of PTSD or mild TBI symptoms. Recent developments in non-pharmacologic sleep treatments have proven efficacious as stand-alone interventions and have potential to improve treatment outcomes by augmenting traditional behavioral and cognitive therapies. This review discusses the extensive scope of work in the area of sleep as it relates to TBI and PTSD, including pathophysiology and neurobiology of sleep; existing and emerging treatment options; as well as methodological issues in sleep measurements for TBI and PTSD. Understanding sleep problems and their role in the development and maintenance of PTSD and TBI symptoms may lead to improvement in overall treatment outcomes while offering a non-stigmatizing entry in mental health services and make current treatments more comprehensive by helping to address a broader spectrum of difficulties.

INTRODUCTION

The prevalence of sleep disorders in the general population (Ohayon, 2007) has gained much attention in recent years, as problems with sleep have been linked to increased psychological sequelae (Vandekerckhove & Cluydts, 2010); increased health problems, particularly cardiovascular health (Lofaso et al., 1996); impaired immune function (Gamaldo, Shaikh, & McArthur, 2012); and even increased mortality (Li, Sato, & Yamaguchi, 2013). The impact of sleep disturbances is now recognized as being wide spread, affecting individuals across age groups, ethnicities, and degrees of health status.

Given the prevalence of sleep difficulties in other neurologic, medical, and psychiatric conditions, it is not surprising that sleep problems are common in individuals with traumatic brain injury (TBI), including those with mild (m)TBI¹. Overall, approximately 40–65% of individuals with mTBI report symptoms of insomnia (Beetar, Guilmette, & Sparadeo, 1996; Dikmen, McLean, & Temkin, 1986). The etiology of these sleep problems is still under investigation, though recent research has implicated neurobiological factors such as impairment in the functioning of neural circuits involved in sleep-wake regulation (Faraguna, Vyazovskiy, Nelson, Tononi, & Cirelli, 2008; Saper, Chou, & Scammell, 2001). Moreover, there is a wide range of sleep-wake dysfunctions that emerge from TBI other than insomnia, including hypersomnia and circadian rhythm abnormalities (Ayalon, Borodkin, Dishon, Kanety, & Dagan, 2007). It is also not clear to what extent restorative sleep is related to successful recovery in individuals with TBI or whether there is a cumulative impact of the severity of TBI on the severity of sleep disorders. What is known is that poor sleep may exacerbate symptoms of TBI (Ouellet & Morin, 2007), impact the individual’s ability to cope with these symptoms (Lew et al., 2009), increase neuropsychiatric symptoms (depression, anxiety and apathy) post-injury (Rao, McCann, Han, Bergey, & Smith, 2014), and inhibit complete participation in rehabilitation treatments (Worthington & Melia, 2006).

PTSD is another major psychiatric disorder with poor sleep being one of its principal symptoms. In PTSD, disturbed sleep is part of the diagnostic criteria, with insomnia being a symptom of alterations in arousal and reactivity (Cluster E), and nightmares as part of the re-experiencing symptom cluster (Cluster B) (American Psychiatric Association, 2013). Consequently, individuals with PTSD tend to present with extremely poor, fragmented, and variable sleep quality as measured by self-report (van Liempt, 2012). As with TBI, it is not clear to what extent the severity of sleep disturbance maintains the psychopathology of PTSD given that poor sleep is associated with increased anxiety, irritability, and deficient coping (K. Babson et al., 2011; Pietrzak et al., 2010; van Liempt, 2012). Recently it has been reported that poor sleep independently predicts the development of depression (Baglioni et al., 2011; Buysse et al., 2008; Jackson, Sztendur, Diamond, Byles, & Bruck, 2014; Peppard, Szklo-Coxe, Hla, & Young, 2006; Szklo-Coxe, Young, Peppard, Finn, & Benca, 2010), PTSD (Bryant, Creamer, O’Donnell, Silove, & McFarlane, 2010; Gehrman et al., 2013; Gerhart, Hall, Russ, Canetti, & Hobfoll, 2014; Kobayashi & Mellman, 2012; Koffel, Polusny, Arbisi, & Erbes, 2013; Spoormaker & Montgomery, 2008; Swinkels, Ulmer, Beckham, Buse, & Calhoun, 2013; Wright et al., 2011), changes in post-deployment depression and PTSD symptoms (Wright et al., 2011), suicide risk in both clinical and non-clinical populations (Pigeon, Pinquart, & Conner, 2012; Ribeiro et al., 2012; and for a review see Woznica, Carney, Kuo, & Moss, 2014), and is more related to low satisfaction with life and post-deployment community integration than other symptoms of PTSD (Kark, Stavitsky, Deluca, Lafleche, & Bogdanova, 2013). Furthermore, there is some evidence that poor sleep likely impacts the efficacy of evidence-based psychotherapy for treatment of PTSD (Nappi, Drummond, & Hall, 2012). This evidence is based on the fact that restorative sleep is necessary for the consolidation of emotional memories (Payne, Chambers, & Kensinger, 2012; Payne & Kensinger, 2011; Stickgold & Walker, 2007; Walker, 2010) and generalization of fear extinction (Pace-Schott et al., 2009; Pace-Schott, Verga, Bennett, & Spencer, 2012) - cognitive processes that are the foundation of most evidence-based treatments for PTSD (van Liempt, 2012).

Given the nature of the recent conflicts, with high rates of physical and psychological injuries due to blast exposures and multiple tours of duty, returning Veterans are often present with co-occurring TBI and PTSD (reviewed in Bogdanova & Verfaellie, 2012). Three of the largest studies evaluating OEF/OIF/OND Veterans report 5–7% of Veterans have probable co-morbid mTBI/PTSD, and in those with mTBI, the rate of co-morbid PTSD is between 33% and 39% (Carlson et al., 2011). Little is known about the overlap or differences between PTSD and TBI with respect to sleep. Few studies have compared individuals with co-occurring PTSD and TBI to those with TBI or PTSD only, to evaluate sleep quality and to determine the differential contribution of these distinct yet overlapping conditions to sleep impairment. One recent study examined PTSD vs. TBI in returning combat Veterans and found no differences between diagnostic groups (PTSD vs. TBI) with the exception of increased arousals among patients with PTSD and greater slow wave sleep (SWS) among those with TBI (Capaldi, Guerrero, & Killgore, 2011). These results suggest some sleep problems in PTSD and TBI might be distinguishable between the two conditions.

The overlap between TBI and PTSD combined with other confounding deployment- and post-deployment-related factors, present significant methodological challenges, which make it difficult to conduct research and to interpret the findings in this population. A recent prospective study explored the role of sleep as a mediating factor in the development of mental health symptoms (TBI, PTSD and depression) in a large sample (29,640) of US Navy and Marine Corps men (Macera, Aralis, Rauh, & MacGregor, 2013). Sleep problems mediated the association between TBI and the development of PTSD and depression. The results suggest that sleep problems may be an early indicator of risk for PTSD or depression in TBI. Further, there is evidence that sleep problems can be considered a separate post-deployment health problem, distinct from PTSD, TBI, and depression (Maguen, Lau, Madden, & Seal, 2012). These and other research findings highlighted in this review, despite their methodological limitations, provide future research directions and suggest that early identification and treatment of sleep disturbance may potentially prevent the development and improve the outcome of mental health disorders, such as PTSD and depression.

This review is intended to 1) provide the latest information on sleep as it relates to the pathophysiology and psychopathology of PTSD and TBI, particularly in military personnel and Veterans returning from more recent Operation Iraqi Freedom/Operation Enduring Freedom/Operation New Dawn (OEF/OIF/OND) conflicts; and 2) to discuss how sleep impacts recovery, course of treatment and outcome in these individuals. Sleep problems are prevalent among OEF/OIF/OND Veterans and are reported to be among the primary complaints (Faestel, Littell, Vitiello, Forsberg, & Littman, 2013; Hoge et al., 2008; McLay, Klam, & Volkert, 2010; Wallace et al., 2011). Veterans returning from combat present with a variety of sleep disturbances that are not limited to insomnia and include circadian rhythm disorders, Obstructive Sleep Apnea (OSA), and even possibly Rapid Eye Movement (REM) Behavior Disorder (Mysliwiec, Gill, et al., 2013; Mysliwiec, McGraw, et al., 2013). Since poor sleep may exacerbate psychological suffering related to combat-related physical (i.e., mTBI) and psychological (i.e., PTSD) trauma and given the evidence that sleep may moderate the efficacy of behavioral treatments for these conditions, it is important to understand the etiology of sleep disorders in this population.

While trauma-focused PTSD treatments have led to some improvements in sleep, nearly half of PTSD treatment positive responders report residual sleep problems, despite relief in other symptom domains (Schoenfeld, Deviva, & Manber, 2012; Zayfert & DeViva, 2004). The emergence of sleep as a distinct post-deployment factor (Maguen et al., 2012) highlights the need to address sleep disturbances in the clinical screening process and to develop sleep-specific treatment protocols. Moreover, sleep treatment has the potential to provide a non-stigmatizing entry into mental health services by focusing on a topic that Veterans may not consider to be ‘mental health treatment’, thereby maximizing engagement in treatment (Epstein, Babcock-Parziale, Herb, Goren, & Bushnell, 2013). For example, a recent case study found that pre-empting PTSD treatment with sleep-focused intervention facilitated successful entry into exposure-based therapy for PTSD (Baddeley & Gros, 2013). Although to date there is a paucity of studies examining the efficacy of sleep treatment for Veterans with mTBI and PTSD, the few recent treatment outcome studies in PTSD have shown a positive impact of sleep treatment on PTSD symptoms (Casement & Swanson, 2012; Germain et al., 2012; Nakamura, Lipschitz, Landward, Kuhn, & West, 2011; Nappi et al., 2012) and may provide a good model for the development of treatment programs for returning Veterans with comorbid TBI and PTSD.

SLEEP DISTURBANCES AND PTSD

Sleep disturbance is one of the core features of PTSD (Calhoun et al., 2007; Dagan, Zinger, & Lavie, 1997). Approximately 70% of patients with PTSD have co-occurring sleep problems (K. A. Babson & Feldner, 2010). PTSD has been associated with reduced sleep efficiency, increased sleep latency, increased sleep fragmentation, and greater night-to-night variability in sleep (Breslau et al., 2004; Calhoun et al., 2007; Habukawa, Uchimura, Maeda, Kotorii, & Maeda, 2007; Mellman, Kulick-Bell, Ashlock, & Nolan, 1995; Straus, Drummond, Nappi, Jenkins, & Norman, 2015). Sleep difficulties in the Veteran population are linked to a variety of poor mental health and socioeconomic outcomes. For example, one recent study found that poor sleep quality was associated with PTSD, panic disorder, depression, suicidal ideation, and risky drinking behavior in Iraq/Afghanistan Veterans (Swinkels et al., 2013). Other studies have reported similar poor outcomes in Veterans with sleep difficulties including lower quality of life and more severe fatigue, pain, PTSD, and depressive symptoms compared to those without sleep difficulties (Lang, Veazey-Morris, & Andrasik, 2014; Wallace et al., 2011). Ribeiro and colleagues (2012) reported an association between insomnia symptoms and suicidal ideation in a sample of 311 military personnel. Specifically, sleep disturbances were found to be predictive of shorter time to suicide (57% loss of survival time) in Veterans in Veterans Health Administration care (Pigeon, Britton, Ilgen, Chapman, & Conner, 2012). Investigating the relationship between sleep disturbances and suicide is important given the high rates of reported sleep disturbances in Veterans Health Administration users (Mustafa, Erokwu, Ebose, & Strohl, 2005) and in returning OEF/OIF/OND Veterans (Hoge et al., 2004; Neylan et al., 1998; Seelig et al., 2010; Wallace et al., 2011); as the rate of suicide is not entirely attributable to a PTSD diagnosis (Lewis, Creamer, & Failla, 2009). This information may help with suicide risk identification and prevention methods as well as early and effective sleep-focused interventions for Veterans at high risk for suicide.

Sleep disturbance itself may be a risk factor for development of PTSD following exposure to trauma (Bryant et al., 2010; Gehrman et al., 2013; Gerhart et al., 2014; Kobayashi & Mellman, 2012; Koffel et al., 2013; Spoormaker & Montgomery, 2008; Swinkels et al., 2013; Wright et al., 2011). A longitudinal investigation of sleep and trauma-related distress found that initial sleep problems predicted increased PTSD and depression at 6 month follow-up, whereas initial PTSD and depression did not predict increased sleep problems (Gerhart et al., 2014). A longer follow- up study similarly found that pre-deployment sleep complaints uniquely and incrementally predicted PTSD and depression up to two years after deployment, even after controlling for baseline psychiatric symptoms and negative emotionality characteristics (Koffel et al., 2013). These findings suggest that pre-trauma sleep disturbance predicts later development of PTSD following exposure to trauma, even after controlling for pre-existing psychopathology. Sleep disturbances may also mediate the effect of mTBI on the development of PTSD and depression and may be an early indicator of risk for depression and PTSD in blast exposed OEF/OIF/OND Veterans (Macera et al., 2013). These findings highlight the importance of pre-deployment sleep symptom screening procedures, which could allow for early identification of Veterans with pre-deployment sleep problems, and facilitate prospective post-deployment care and health planning.

Sleep problems are particularly common in returning Veterans with PTSD compared to those without PTSD or trauma history (N. Orr et al., 2010). A recent review of deployment-related insomnia in military personnel provides a thorough overview of how sleep problems develop in military personnel and Veterans and are perpetuated by the deployment experience (Bramoweth & Germain, 2013). Mellman and colleagues (1995) evaluated sleep quality among combat Veterans with PTSD and found that the most commonly reported sleep-related symptoms were recurrent awakenings, threatening dreams, thrashing movements while sleeping, and waking up startled or in a panic. One recent study by Mysliwiec and colleagues (2013) found an increased rate of comorbid diagnoses of insomnia and OSA in US military personnel returning from deployment presenting to a sleep clinic. In this retrospective cross-sectional study sleep disorders were diagnosed in 88.2% of subjects with 62.7% meeting diagnostic criteria for OSA and 63.6% for insomnia. Those with comorbid insomnia and OSA were more likely to meet criteria for PTSD and depression (Mysliwiec, McGraw, et al., 2013). Wallace and colleagues (2011) characterized sleep quality in OEF/OEF/OND Veterans with PTSD and insomnia (with and without mTBI). Fifty-six percent of the PTSD group reported severe insomnia and 44% reported moderate insomnia. Both groups (PTSD vs. mTBI) reported severe difficulties with sleep initiation, maintenance, early awakenings, and daytime fatigue, while daytime sleepiness was greater in the PTSD-mTBI group. When compared to the PTSD-mTBI insomnia group, the PTSD group showed worse sleep continuity measured by actigraphy (ACT) (sleep efficiency, nighttime awakenings, and fragmentation index). These data suggest that, while both groups reported similar insomnia severity, patients with PTSD might be experiencing more sleep quality difficulties while the PTSD-mTBI patients might be suffering from a differential set of difficulties resulting from a complex interplay between insomnia and hypersomnia.

Sleep Discrepancies in PTSD: Current Issues

While subjective sleep complaints are a core feature in PTSD, previous studies have been unable to consistently corroborate subjective sleep complaints (i.e., self-report questionairres, sleep diaries) with objective sleep measurements (i.e., polysomnography [PSG], ACT) (Germain, Hall, Shear, Nofzinger, & Buysse, 2006; Pillar, Malhotra, & Lavie, 2000; Woodward, Bliwise, Friedman, & Gusman, 1996). One issue might be that PSG studies typically rely on only 1 or 2 nights of data collection, which is problematic if symptom severity is intermittent or highly variable. Indeed, recent evidence suggests greater intraindividual variability of sleep in military-related PTSD, relative to primary insomnia patients and healthy controls (Straus et al., 2015). Further, it has been suggested that many patients with PTSD may sleep better in a PSG laboratory because they feel safer or more protected under the observation of research staff compared to their home environment, which compromises the ecological validity of the data (Germain et al., 2006; Khawaja, Hashmi, Aftab, Westermeyer, & Hurwitz, 2014). In line with this theory, results from a small pilot study revealed a relationship between in-home PSG-measured sleep disturbances and subjective sleep problems in a sample of adult crime victims with PTSD (Germain et al., 2006). While PSG is the gold standard for objective sleep measurement, it is also costly and might not be the most practical or ecologically-valid method to examine sleep in the PTSD population.

While the ambulatory nature of actigraphic monitoring allows for more ecologically-valid data collection and prolonged sleep recording compared to PSG, the relationship between ACT and self-reported sleep quality also remains equivocal (for a review see Khawaja et al., 2014). For example, prior work reported a significant relationship between sleep diary and ACT total sleep time (TST), but sleep diaries underestimated the number of sleep awakenings compared to ACT (Westermeyer et al., 2007). These findings suggest that ACT and diary data may align on some but not all measures of sleep quality. In a recent CBT-I intervention study for chronic PTSD that used both ambulatory PSG and ACT, the treatment group showed post-treatment improvements in both PSG-measured TST and self-reported sleep symptoms, compared to the waitlist group (Talbot et al., 2014). The PSG data were only based on one night of data collection and there were no significant improvements in continuous ACT-measured sleep to corroborate the PSG data or self-reported sleep improvements. More research is necessary to understand the accuracy of PSG and ACT-measured sleep in the PTSD population.

Although the disparities between objective and subjective sleep data are often considered a source of measurement error, it has also been suggested that the difference between objective and subjective sleep measurements, or sleep discrepancies, might capture clinically meaningful correlates of psychiatric status (Williams, Kay, Rowe, & McCrae, 2013). Prior work has been unable to demonstrate a relationship between sleep discrepancies and PTSD status (Kobayashi, Huntley, Lavela, & Mellman, 2012), which is thought to reflect sleep misperceptions in PTSD (Dagan et al., 1997; Klein, Koren, Arnon, & Lavie, 2003; Kobayashi et al., 2012). Indeed, “paradoxical insomnia”-a relatively new diagnosis that first appeared in the International Classification of Sleep Disorders - 3rd Edition (ICSD-3; American Academy of Sleep Medicine, 2014)- is characterized by quantifiable discrepancies between objective and subjective sleep measurements. While paradoxical insomnia is relatively rare amongst civilians with insomnia (5%), a recent study reported paradoxical insomnia in 17.2% of a large cohort of active duty military personnel with sleep disorders (Mysliwiec, McGraw, et al., 2013). Further, results of a recent study reported underestimated self-reported TST compared to ACT TST, which offers new evidence for PTSD-related paradoxical insomnia (Ghadami, Khaledi-Paveh, Nasouri, & Khazaie, 2014). These findings highlight the importance of utilizing objective sleep measures in the military population, without which paradoxical insomnia would go undetected.

Despite the prevalence of paradoxical insomnia, individuals with PTSD do likely suffer from objective insomnia. However, the experience of true insomnia related to acute PTSD may give rise to enduring thought patterns about sleep that eventually manifest as subjective insomnia in chronic PTSD (Dagan et al., 1997). A recent PSG study demonstrated fragmented and reduced REM sleep in PTSD patients with relatively proximate trauma exposure compared to unaltered REM sleep in the chronic or ‘distal’ PTSD patients, even after controlling for depression (Mellman, Kobayashi, Lavela, Wilson, & Hall Brown, 2014). Sleep state misperceptions in paradoxical insomnia might be especially common for returning Veterans as they reintegrate into a safe sleeping environment after maintaining high levels of nocturnal vigilance during deployment (Mysliwiec, McGraw, et al., 2013). Further, a recent neuroimaging study reported a link between decreased subjective sleep quality and reduced cortical and frontal grey matter volume in Gulf War Veterans without a current psychiatric diagnosis, which may suggest an association between frontal lobe integrity and the capacity to experience sleep as restorative (Chao, Mohlenhoff, Weiner, & Neylan, 2014). These findings provide potential neural evidence to corroborate the theory of dysfunctional cognitions and thought patterns about subjective sleep quality. Further work is necessary to 1) understand the prevalence and clinical implications of sleep discrepancies or ‘paradoxical insomnia’ in returning Veterans with and without PTSD, 2) investigate if these discrepancies arise during the shift from acute (or ‘proximal’) PTSD to chronic (or ‘distal’ PTSD), and 3) develop early and effective interventions to alleviate sleep misperceptions that might arise in military-related chronic PTSD.

Neurobiological Correlates of Sleep in PTSD

Neuroimaging studies are beginning to explore the brain-behavior link in individuals with PTSD and have recently reported neurobehavioral correlates of sleep disturbance in those with PTSD. Structural and functional neuroimaging studies have found hyperactivation of the amygdala and impaired medial prefrontal cortex function in patients with PTSD (Germain, Buysse, & Nofzinger, 2008; Hughes & Shin, 2011). Specifically, patients with PTSD show hyperactivation of the amygdala in response to threat-related stimuli, which exaggerates the fear response (Germain et al., 2008; Hughes & Shin, 2011; Patel, Spreng, Shin, & Girard, 2012). Activation of the medial prefrontal cortex, which is critically involved in fear extinction, might be impaired in PTSD subjects compared to trauma-exposed controls (Liberzon & Sripada, 2008). With regard to sleep, investigators have found both hypo and hyperactivation of specific brain regions in individuals with PTSD during sleep. For example, Germain and colleagues (2013) found that OEF/OIF/OND Veterans with PTSD experience elevated neural activity during wakefulness and REM sleep in the REM-generating and arousal regions of the brainstem. Results showed hyperactivation of the brain in arousal-related regions involved in REM sleep regulation (midbrain reticular formation left and right locus coeruleus, raphe nuclei, pedunculopontine, laterodorsal tegmental nuclei, cerebellum) as well as limbic structures involved in emotion and memory processing (hippocampus, amygdala, basal ganglia, insula, and thalamus).

Nightmares in PTSD may be facilitated by the hypoactivity of the dorsal medial prefrontal cortex, which modulates the activity in subcortical limbic regions, such as the amygdala, during REM sleep (Germain et al., 2008; Germain et al., 2013). In a similar vein, earlier studies found that symptomatic awakenings in those with PTSD were preceded by REM sleep, suggesting that arousal may intrude upon the sleep state via increased amygdala activity or impaired medial prefrontal cortex function (Mellman, Bustamante, Fins, Pigeon, & Nolan, 2002; Mellman et al., 1995). An electroencephalographic study by Cohen and colleagues (2013) compared sleep architecture between combat Veterans with and without PTSD and found no difference between measures of central arousal (beta and gamma activity) between groups. The authors suggest that central arousal might be indicative of primary insomnia whereas hyperarousal symptoms in PTSD might reflect increased autonomic and limbic activity and/or dysfunction in sleep-related memory or emotional processing. The results from this study also revealed that, in combat-exposed Veterans without PTSD, REM and non-rapid eye movement (NREM) sleep sigma activity was positively correlated with level of combat exposure and NREM sigma activity was negatively correlated with PTSD avoidance symptoms. The authors concluded that these data suggest the possibility of a sleep architecture-related marker of resilience to PTSD (Cohen et al., 2013).

In 2013, van Liempt and colleagues reported the results of a study investigating Hypothalamic-Pituitary-Adrenal (HPA) axis and sleep dysfunction in Veterans with PTSD compared to trauma-exposed and healthy controls (van Liempt et al., 2013). The group collected objective and subjective measures of sleep quality as well as cortisol, adrenocorticotropic hormone (ACTH), and melatonin levels. While no difference in sleep architecture was found between the 3 groups, the PTSD group experienced more nighttime awakenings, which were positively correlated with ACTH levels. ACTH and cortisol were also inversely related to slow wave sleep and ACTH was found to be an independent predictor of SWS in those with PTSD. The authors suggest that fragmented sleep in PTSD might be associated with increased corticotrophin-releasing hormone, which inhibits SWS. The authors did not find evidence of phase shifts in circadian rhythms of cortisol, ACTH, or melatonin in Veterans with PTSD. Of note, a recent study by the same group found a strong relationship between sleep fragmentation and blunted growth hormone secretion in PTSD, both of which might negatively impact sleep-dependent memory consolidation (van Liempt, Vermetten, Lentjes, Arends, & Westenberg, 2011). Together, the foregoing studies provide evidence for dysregulation of arousal centers of the brain related to sleep disturbances in PTSD.

The HPA axis is a neurohormonal feedback loop that functions as part of the stress response system and regulates other biological processes including mood and emotions, digestion, and physical activity. Among the more important and most easily measurable hormones released through the HPA axis is cortisol, which can serve as an index of HPA activity and marker of circadian rhythms; the release of which is activated by the feedback loop initiated by the corticotrophin-releasing hormone secreted by the paraventricular nucleus of the hypothalamus. The release of corticotrophin-releasing hormone from the hypothalamus is influenced by the sleep/wake cycle in addition to having an endogenous circadian rhythm. Studies have found that upon awakening, there is an immediate rise in cortisol, known as the cortisol awakening response (CAR), which reaches its peak approximately 30 minutes following awakening (Fries, Dettenborn, & Kirschbaum, 2009) and demonstrates gradual decline throughout the day (Clow, Thorn, Evans, & Hucklebridge, 2004; Pruessner et al., 1997; Wilhelm, Born, Kudielka, Schlotz, & Wust, 2007). This response is sensitive to a variety of inter-individual differences, including stress, health status, awakening time, and the morningness/eveningness of the individual (Bailey & Heitkemper, 2001; Brigitte M. Kudielka, Buchtal, Uhde, & Wust, 2007; B. M. Kudielka, Federenko, Hellhammer, & Wust, 2006; Pruessner et al., 1997; Randler & Schaal, 2010; Wust, Federenko, Hellhammer, & Kirschbaum, 2000). However, despite these inter-individual differences, the CAR demonstrates within-individual stability (Hucklebridge, Hussain, Evans, & Clow, 2005; Pruessner et al., 1997; Wust et al., 2000).

In so far as the cortisol response and associated HPA axis activity is an intrinsic part of circadian rhythms and is associated with sleep wake activity, it is expected that in individuals with disturbed sleep or altered sleep-wake rhythms these mechanisms would demonstrate significant modulation. However, the results of the studies examining the relationship between the CAR and sleep efficiency have been inconsistent. Some studies have reported no difference between subjective measures of sleep quality and cortisol at awakening (Zhang et al., 2011), while others have found a relationship between sleep duration and sleep disturbances and morning cortisol levels (Backhaus, Junghanns, & Hohagen, 2004; Kumari et al., 2009). Others have reported an association between poor sleep and evening cortisol levels (Rodenbeck, Huether, Ruther, & Hajak, 2002; A. N. Vgontzas et al., 2001). The implication of these equivocal findings is that it is likely that the inter-individual differences and the methods of measuring cortisol (i.e., CAR, diurnal cortisol slope) may result in inconsistent results. Moreover, the majority of these studies measure sleep quality by using conventional sleep measures such as sleep efficiency, duration, and sleep onset latency.

Dysfunction of the HPA axis has been repeatedly implicated in PTSD (de Kloet et al., 2006; Drake, Roehrs, & Roth, 2003). Particularly, the suppression of the HPA axis in PTSD has been reported, with less variability (fewer peaks and troughs) in the diurnal cortisol cycle in individuals with PTSD (Yehuda, 2002). Recently, Wahbeh and Oken (2013) reported lower cortisol across the entire day (wake, 30 minutes after waking, and bedtime) in combat Veterans with PTSD compared to combat Veterans without PTSD. Patients who develop PTSD following exposure to trauma show a sharp increase of cortisol immediately following the traumatic event followed by tapering to low basal cortisol levels over the course of a few months into the chronic phase of PTSD (Pervanidou & Chrousos, 2010). Pervanidou and Chrousos (2010) have posited that the initial spike in cortisol and subsequent low basal cortisol state in the presence of progressive norepinephrine levels characterizes the development and maintenance of PTSD, respectively. In concordance with these findings, it has been reported that symptom severity in chronic PTSD is negatively correlated with cortisol levels (Olff, Guzelcan, de Vries, Assies, & Gersons, 2006). Future work is needed to elucidate the role of HPA response in the development and maintenance of PTSD in returning service members, with the possibility that HPA function might interact with TBI status.

Prior work has also found reduced plasma brain-derived neurotrophic factor (BDNF)—a neuroprotective growth factor that directly influences sleep need (Faraguna et al., 2008)—in PTSD patients, relative to healthy controls (Dell’Osso et al., 2009). Dysregulation of BDNF might adversely affect neuroplasticity, cognition, and memory in PTSD (Deppermann, Storchak, Fallgatter, & Ehlis, 2014; Kaplan, Vasterling, & Vedak, 2010). Chronic stress and cortisol elevation are linked with BDNF down-regulation in regions of the brain such as the ventromedial prefrontal cortex (Gourley, Kedves, Olausson, & Taylor, 2009) and the hippocampus (Rasmusson, Shi, & Duman, 2002)—regions critical in the regulation of physiological arousal and memory, respectively. Re-exposure to a fear-conditioned context reduces BDNF expression in the hippocampus, which could possibly hamper hippocampal-based fear extinction learning (Rasmusson et al., 2002). These results demonstrate a link between BDNF, stress, and cognition. In regards to sleep, animal work has demonstrated increased wakeful daytime exploration is associated with increased cortical BDNF expression and increased slow-wave activity during subsequent sleep, compared to rats awake for the same amount of time but not exploring (Huber, Tononi, & Cirelli, 2007). BDNF regulation seems to be important in the link between daytime cognition and activity with the normal drive for sleep. Future work is needed to disentangle the possible interaction between BDNF regulation, sleep drive and intensity, and cognition in PTSD.

Conclusion

Sleep problems are prevalent in returning Veterans with PTSD. Sleep problems pose a risk factor for the development and maintenance of PTSD and have substantial impact on psychosocial functioning and recovery from PTSD. Furthermore, impaired sleep might also impede upon the effectiveness of clinical treatments for PTSD, the mechanisms for which will be further elaborated on in the following section. Sleep neuroimaging studies have demonstrated regional activation differences between individuals with and without PTSD and have demonstrated increased limbic activity during sleep, which may serve as a marker for nightmares and increased arousal in PTSD that is different in patterns of arousal in insomnia. Results of studies investigating the possible link between HPA-axis dysfunction and sleep in PTSD have been largely equivocal, although the link between both PTSD and sleep-related HPA-axis circadian activity is well established. New evidence suggests a link between BDNF function, sleep, and cognition in PTSD, but this emerging line of inquiry remains open. Research on sleep problems in OEF/OIF/OND Veterans, especially those with PTSD, is beginning to explore the complexity and correlates of these coexisting disorders. However, there is more work to be done as it is imperative that we hone sensitive sleep measures in order to elucidate the neural and behavioral connection between sleep disturbance and PTSD development and maintenance, and the impact of sleep on outcomes in PTSD.

Impact of poor sleep on cognition in PTSD

Individuals with PTSD often complain of memory difficulties that have been hypothesized to be related to hippocampal dysfunction in these individuals (Bremner, 2007). The work of Vasterling and colleagues (2002) supports the association between PTSD and neurocognitive and intellectual performance deficits, citing increased problems with learning, memory, and attention in Vietnam Veterans with PTSD. Other studies have similarly found that in other cohorts of individuals with PTSD, difficulties in attention, learning, and verbal memory are also present (Bremner et al., 1995; Marx, Doron-Lamarca, Proctor, & Vasterling, 2009; Samuelson et al., 2006; Vasterling et al., 2012; Vasterling et al., 2002). Few studies have examined the circumscribed relation between sleep and cognitive function in individuals with PTSD. The potential role of sleep disturbance on PTSD has been peripherally addressed previously in work examining the role of sleep in learning and emotional memory consolidation (Nishida, Pearsall, Buckner, & Walker, 2009; Stickgold, Hobson, Fosse, & Fosse, 2001; Stickgold & Walker, 2005) or by demonstrating that even short periods of sleep such as daytime naps can diminish impulse control (i.e., reactivity to anger and fear inducing stimuli (Gujar, Yoo, Hu, & Walker, 2011). Yoo, Gujar, Hu, Jolesz and Walker (2007) demonstrated the mechanisms for how disturbed sleep can exacerbate and maintain PTSD symptoms by showing that sleep deprivation can enhance negative emotional response by disrupting connectivity between the amygdala and the medial prefrontal cortices, a pathway that modulates emotional responsivity. However, few studies have examined specific components of cognitive difficulties in PTSD, including attention, executive function, and learning, with respect to poor sleep. One recent study by Brownlow, Brown, and Mellman (2014) demonstrated a relationship between sustained attention, verbal memory, and objectively measured sleep in PTSD. Interestingly, a full night’s sleep (approximately 7 hours or greater) mitigated attention errors in PTSD.

Sleep problems and associated cognitive deficits may have an impact on treatment outcome in PTSD. Evidence-based treatments for PTSD, such as cognitive processing therapy or prolonged exposure treatment, have strong learning components involving attention and memory consolidation processes. Furthermore, extinction of conditioned fear responses, which is related to consolidation of emotional memories, is also an important part of psychotherapeutic interventions aimed at alleviating PTSD. Individuals with PTSD have impaired extinction of conditioned fear responses and impaired recall of the extinction (Blechert, Michael, Vriends, Margraf, & Wilhelm, 2007; Milad et al., 2008; S. P. Orr et al., 2000; Wiseman-Hakes et al., 2013). Given the evidence that sleep—and particularly REM sleep—plays an important role in memory consolidation, including that of emotional memories (Bennion, Mickley Steinmetz, Kensinger, & Payne, 2013; Hu, Stylos-Allan, & Walker, 2006; Payne et al., 2012; U. Wagner, Gais, & Born, 2001; U Wagner, Hallschmid, Rasch, & Born, 2006), and that sleep promotes generalization of extinction memory (Pace-Schott et al., 2009), it is important to consider the impact of poor sleep on treatment outcome in individuals with PTSD.

Conclusion

Overall, the impact of sleep on cognitive functioning is well studied, although due to the multiple etiologies of poor sleep and distinct methodologies of studying sleep, the conclusion of whether poor sleep has a long-term impact on cognitive function is equivocal. Additionally, the neurobiological mechanism behind the interaction between poor sleep and cognition has yet to be elucidated. However, what is clear from a multitude of studies is that there is an association between sleep and daytime cognition and consolidation of memories, particularly emotionally salient ones. Individuals impacted by PTSD are already vulnerable to cognitive deficits and are at high risk for poor sleep. Therefore, in individuals with PTSD, the additional impact of poor sleep on their everyday function coupled with the loss of benefits associated with restorative sleep is particularly burdensome. In addition to impacting their everyday function, poor sleep may also have a negative effect on psychotherapeutic treatment in those with PTSD.

Management and Treatment of Sleep Problems

Pharmacologic treatments

A thorough review of the pharmacologic treatments of sleep problems is beyond the scope of this paper and was recently reviewed by Richey and Krystal (2011). Briefly, the current first line of pharmacologic treatment for insomnia are the benzodiazepine-receptor agonists medications: zolpidem, zaleplon, and eszopiclone. The main advantage of these medications over the traditional benzodiazepine drugs is their relatively shorter duration, fewer side effects, and low abuse potential (Richey & Krystal, 2011). However, despite the general improvement in efficacy these medications are not consistently effective and have significant side effects including headaches, anxiety, and sleepiness (Krystal, Erman, Zammit, Soubrane, & Roth, 2008). For example, the most common prescribed of these, zolpidem, has not demonstrated consistent effectiveness in sleep maintenance, but is generally helpful in sleep initiation (Rosenberg, 2006). Other pharmacologic treatments for insomnia have included tricyclic antidepressants as well as serotonergic agents such as trazodone, but most medications have not been thoroughly investigated in randomized controlled trials and can have significant side effects.

Prazosin is an alpha-1 antagonist that has been used pharmacologically to treat PTSD-related sleep disturbances and nightmares. Prazosin’s mechanism of action has been stipulated to be the attenuation of noradrenergic function during REM sleep, thereby reducing nightmares and related complaints of insomnia (Raskind et al., 2003; Ross, Gresch, Ball, Sanford, & Morrison, 1995). Studies have shown that prazosin-related improvement in sleep is accompanied by improvements in other daytime symptoms of PTSD (Boynton, Bentley, Strachan, Barbato, & Raskind, 2009; Germain et al., 2012; Raskind et al., 2007; Raskind et al., 2003; Raskind et al., 2002; Taylor & Raskind, 2002). A recent randomized control trial of prazosin in returning Veterans with PTSD showed favorable results with reported improvements in nightmares, overall sleep quality, global function, and PTSD symptoms (Raskind et al., 2013). However, substantial residual symptoms remained, suggesting that adjunct treatment was needed.

In addition to prescription medication, over-the-counter pharmacologic sleep enhancers include melatonin, which is a naturally occurring hormone that is endogenously produced by the pineal gland and is important in regulation of circadian rhythms by initiating the feedback circuitry of the suprachiasmatic nucleus. Melatonin has been explored as a hypnotic, but a number of randomized trials have failed to find evidence of efficacy (Richey & Krystal, 2011). Overall, although short-term effects of pharmacotherapy on sleep problems and particularly insomnia are well documented, the long-term benefits are unknown, and long-term use may be associated with multiple adverse effects (Gillin & Byerley, 1990). Some have suggested that pharmacologic sleep agents may negatively impact neural recovery/neuroplasticity in patients recovering from TBI (Larson & Zollman, 2010). Moreover, it has been recently suggested that biological treatments might be the most efficacious for short sleep duration insomnia, while insomnia with normal sleep duration might respond better to psychological treatment (A. Vgontzas & Fernandez-Mendoza, 2013). While some antidepressants prescribed for PTSD might activate BDNF pathways, further research is needed to develop agents that selectively target BDNF function, which have the potential to alleviate cognitive and sleep symptoms in PTSD and TBI (Kaplan et al., 2010). More clinical studies are necessary to understand the efficacy of BDNF-enhancing interventions, such as exercise, and combined pharmacological-behavioral approaches (Chen, Ivy, & Russo-Neustadt, 2006; Kaplan et al., 2010).

Non-pharmacologic treatments

The treatment and management of sleep disturbance has primarily utilized pharmacologic approaches that have the potential for side effects and, in some medications, development of tolerance over time. As sleep disturbances occur for a variety of reasons, including mood difficulties, behavioral factors such as poor stimulus control, aging-related changes, pain, and other medical conditions, effective and consistent pharmacologic treatment has been difficult to achieve. There is also a growing body of evidence for effectiveness of non-pharmacological interventions, such as cognitive and behavioral treatments, for sleep disturbance in PTSD (Schoenfeld et al., 2012). Given the complexity and multitude of underlying psychological factors that commonly accompany sleep disturbance (Espie et al., 2012), including behavioral intervention in the treatment plan may be beneficial for patients with comorbid PTSD and mTBI (polytrauma for the Veteran population). The current practice guidelines developed by the Standards of Practice Committee of the American Academy of Sleep Medicine recommend using psychological and behavioral interventions for the treatment of chronic primary insomnia (Standard) and secondary insomnia (Guideline) (Morgenthaler et al., 2007). As of 2012, no guidelines specific to treatment of sleep disorders following mTBI have been established (Marshall et al., 2012). Recommendations for sleep disorders have been adopted from current practice guidelines outside of the TBI field. Practice guidelines for incorporation of sleep interventions into management of PTSD are yet to be established (Germain et al., 2013), however, the PTSD algorithm of the Psychopharmacology Algorithm Project at the Harvard South Shore Program has proposed treatment guidelines outlining that sleep evaluation and treatment should be the first step in assessing and treating PTSD (Bajor, Ticlea, & Osser, 2011).

In terms of non-pharmacologic treatments, psychotherapeutic interventions for insomnia have gained momentum in the recent years with the increase of evidence for their efficacy. Cognitive behavioral treatment for insomnia (CBT-I) is a psychotherapeutic treatment that has shown efficacy in treating poor sleep particularly in insomnia (Edinger, Wohlgemuth, Radtke, Marsh, & Quillian, 2001; Galuszko-Węgielnik, Jakuszkowiak-Wojten, Wiglusz, Cubala, & Landowski, 2012; Morin et al., 2006) by targeting maladaptive thoughts and behaviors associated with maintaining the cycle of poor sleep. Specifically, the goal of CBT-I is to target factors that maintain insomnia over time including sleep-related anxiety, maladaptive or sleep-interfering behaviors, and the resultant dysregulated sleep drive (Mitchell, Gehrman, Perlis, & Umscheid, 2012). CBT-I uses stimulus control, sleep restriction, and cognitive restructuring to change sleep-related behaviors and thoughts that maintain poor sleep. The treatment is usually administered over the course of 4–8 sessions, with significant improvements often around sessions 3 or 4 (Edinger & Means, 2005; Mitchell et al., 2012). Overall, randomized control trials have found that CBT-I has medium to large effects in reducing insomnia, in comparison to placebo treatments such as wait-list and sleep hygiene programs, that are maintained post-treatment (Sanchez-Ortuno & Edinger, 2012), but unstandardized treatment adherence across protocols complicate the evaluation of CBT-I treatment outcome (Matthews, Arnedt, McCarthy, Cuddihy, & Aloia, 2013). Recently, the favorable efficacy of CBT-I in Veterans and implementation feasibility by VA providers has already been reported (Karlin, Trockel, Taylor, Gimeno, & Manber, 2013). In comparison to pharmacologic treatments trials compared CBT-I to zopiclone, zolpidem, temazepam, and triazolam and overall, found that CBT-I is at least as effective in treating insomnia when compared to these medications (Mitchell et al., 2012). Furthermore, CBT-I efficacy is likely more durable than that of medication (Mitchell et al., 2012).

In the Veteran population, CBT-I has been used to treat PTSD-related sleep disturbances. In a multi-component treatment study, those with PTSD demonstrated post-treatment improvements in their sleep onset latency, awakenings, sleep efficiency, and self-reported sleep quality in response to CBT-I (Ulmer, Edinger, & Calhoun, 2011). However, the nature of the sleep disturbances in individuals with PTSD is substantially different than those experienced by other individuals with sleep problems (see the PTSD and sleep section) and consequently some investigators have aimed at tailoring sleep treatments to sleep disturbances experienced by those with PTSD.

Imagery rehearsal therapies (IRT) are aimed at modulating post-traumatic nightmares by decreasing their frequency and severity through “rescripting” of nightmare content (Kellner, Neidhardt, Krakow, & Pathak, 1992; Kellner, Singh, & Irigoyen-Rascon, 1991) and were first shown to be an effective treatment for nightmares in individuals with PTSD by Krakow and colleagues (2001). There are several different protocols for imagery rehearsal in the PTSD population including Exposure, Relaxation, and Rescripting Therapy (Davis & Wright, 2006), Imagery Rehearsal Therapy (Krakow & Zadra, 2006), and Imagery Rehearsal and Exposure Therapy (Long et al., 2011). Regardless of specific protocol, the main components of each of these treatments are similar and include: psychoeducation, narratives about selected nightmares, and imaginal rehearsal of new dreams, as well as some forms of exposure to the selected nightmares. Over the years, empirical evidence for the efficacy of IRT has been growing and has been shown to reduce nightmare frequency in US Army Veterans (Moore & Krakow, 2007) and female sexual assault victims (Krakow et al., 2001). A recent meta-analysis confirmed that IRT reduces nightmare frequency and improves sleep quality in a variety of trauma-related study samples and protocols (Casement & Swanson, 2012). On the contrary, the largest study to date, conducted in Vietnam era Veterans did not find IRT to be effective, compared to an active control condition (Cook et al., 2010) therefore, the efficacy of IRT in Veterans with PTSD is still not fully determined. Further, a recent review of the current evidence base for IRT in PTSD revealed a lack of higher-level study designs and methodological inconsistencies across treatment protocols, none of which have been compared directly (Harb et al., 2013). These studies highlight the limited data supporting the efficacy of IRT for the treatment of PTSD.

Others have combined IRT and CBT-I treatments or components of each of these cognitive behavioral therapies in order to target both the PTSD-related nightmares as well as the general sleep dysregulation that is especially prevalent in the Veteran population. Studies of combined treatment have varied from group treatments to individual therapies and have found significant effect sizes in improvement of nightmares, sleep quality, and PTSD symptoms (Davis & Wright, 2006; Germain, Shear, Hall, & Buysse, 2007; Krakow et al., 2001; Nappi et al., 2012; Ulmer et al., 2011). Overall, these studies demonstrate a greater improvement in overall sleep quality than IRT alone (Casement & Swanson, 2012). Furthermore, at least one study comparing a behavioral intervention for sleep and nightmares found that the behavioral treatment was comparable to prazosin in reducing insomnia severity and PTSD-related daytime symptoms in Veterans with PTSD (Germain et al., 2012). Another recent study reported improved subjectively and objectively measured sleep and a reduction in PTSD symptom severity and PTSD-related nighttime symptoms in OEF/OIF/OND combat Veterans with PTSD following 4 sessions of combined CBI-I/IRT, compared to a waitlist control group (Margolies, Rybarczyk, Vrana, Leszczyszyn, & Lynch, 2013). A recent case report described that treatment of insomnia in a Veteran patient with CBT-I improved sleep and facilitated entry into exposure therapy for PTSD (Baddeley & Gros, 2013). The preliminary results of this case study also highlight the importance of investigating the most efficacious order of evidence-based psychotherapies, particularly in comorbid insomnia and PTSD.

Conclusion

Treatment of sleep disorders has relied on pharmacologic management for many years. More recently other methods of treating sleep disorders have gained attention, particularly with sleep problems gaining recognition for their prevalence and the long-term efficacy and side effects of pharmacologic agents remaining in question. Behavioral treatments of insomnia are promising, particularly for individuals with insomnia related to mood and anxiety disorders including individuals with PTSD. For Veterans returning from the OEF/OIF/OND arenas, sleep problems are especially prevalent and are often co-occurring with both PTSD and mTBI (Wallace et al., 2011). These young Veterans are particularly encumbered by the mood, cognitive, and health difficulties that are associated with PTSD and without the additional burden of sleep problems. Behavioral treatments of PTSD do not address sleep as a primary target, yet sleep difficulties may be an important barrier for good treatment outcome (Nappi et al., 2012). This area of research is currently lacking in the literature, particularly as it pertains to PTSD.

SLEEP DISTURBANCES FOLLOWING TRAUMATIC BRAIN INJURY

Sleep and circadian disturbances in TBI

Daytime sleepiness and nighttime sleep disturbance are common following TBI (Baumann, 2012). Common post-traumatic sleep-wake disturbances (SWDs) include; poor nocturnal sleep quality, hypersomnia, excessive daytime sleepiness, impaired daytime vigilance, and circadian rhythm phase shift (Ayalon et al., 2007; Baumann, 2012; Baumann, Werth, Stocker, Ludwig, & Bassetti, 2007; Kempf, Werth, Kaiser, Bassetti, & Baumann, 2010; Ponsford, Parcell, Sinclair, Roper, & Rajaratnam, 2013). Objective sleep studies have shown that patients with TBI exhibit reduced sleep efficiency, increased wakefulness after sleep onset, increased sleep onset latency, alterations in sleep architecture (less REM sleep, more time in “light” NREM sleep), and report more frequent insomnia (Baumann et al., 2007; Collen, Orr, Lettieri, Carter, & Holley, 2012; Mathias & Alvaro, 2012; Ponsford et al., 2012; Shekleton et al., 2010). Reports have shown that 25–81% of post-acute TBI patients suffer protracted or even permanent SWDs or other sleep disturbances (Mathias & Alvaro, 2012), regardless of localization, severity of, or time since, the TBI (Baumann et al., 2007; Castriotta et al., 2007; Kempf et al., 2010). Not only is the prevalence of sleep problems in TBI patients significantly higher than in the general population (Mathias & Alvaro, 2012), these sleep problems are chronic and have been shown to persist at 6 months and up to 5 years post-injury (Beetar et al., 1996; Fogelberg, Hoffman, Dikmen, Temkin, & Bell, 2012; Kaufman et al., 2001; Kempf et al., 2010; Masson et al., 1996; Stulemeijer et al., 2006; Tham et al., 2012; Watson, Dikmen, Machamer, Doherty, & Temkin, 2007).

There is a lack of understanding regarding the temporal relationship between primary brain injury and the subsequent emergence and maintenance of SWDs. SWDs are common during the acute TBI phase, as evidenced by objectively and subjectively measured hypersomnia and reduced sleep efficiency (Chiu, Lo, Chiang, & Tsai, 2014) and can persist up to 5 years post-injury (Masson et al., 1996). However, many of the long-term follow-up studies are based on subjective sleep measures, thus, it is possible that protracted sleep complaints are related to sleep state misperceptions following the initial experience of objective sleep problems (Orff, Ayalon, & Drummond, 2009). Prior work has also demonstrated that the spectrum of mild to severe TBI patients is affected by objectively measured chronic SWDs (Verma, Anand, & Verma, 2007). The type of brain injury (e.g., diffuse, focal) might influence the type and severity of SWD (Orff et al., 2009). Recent animal work has shown hypersomnia during the first week following induction of diffuse TBI, however post-traumatic hypersomnia did not develop into chronic sleep disturbance (Rowe, Harrison, O’Hara, & Lifshitz, 2014). These data suggest that hypersomnia following diffuse TBI (e.g., blast exposure) does not necessarily develop into persistent sleep problems. A recent study by Ponsford and colleagues (2013) demonstrated that poor sleep quality and daytime sleepiness increases with time since injury in older TBI patients relative to younger TBI patients, suggesting an interaction between sleep disturbances, aging, and time since injury in TBI. These findings highlight the importance of considering the course of sleep disturbances in TBI in Veterans across the life span as Veterans age. Future work is needed to understand the relationship between injury severity and type and the development and persistence of SWDs.

A recent meta-analysis using data from 21 studies (n=1706) examined twenty-one years of TBI and sleep literature (1990–2011) (Mathias & Alvaro, 2012). The group reported that in studies that used objective sleep measures in conjunction with formal diagnostic criteria; the prevalence of sleep disturbances was 53%. Compared to the prevalence reported in the general population, the TBI group was found to be significantly more likely to experience sleep disturbances and was more than 12 times more likely to have diagnosed sleep disorders. The authors also reported several challenges to the clinical utility of these studies due to differences in criteria, which may explain the wide range of sleep disturbance incidences described in the literature. In similar vein, a survey by the Armed Forces Health Surveillance System found that sleep disorders were among the most common sequelae following TBI in the active duty Armed Forces service members between the years 2000–2012 (AFHSC, 2013). Farrell-Carnahan, Franke, Graham, and McNamee (2013) recently found that in a sample of 114 returning Veterans with blast-induced mild traumatic brain injury, 77% reported subjective sleep difficulties. The foregoing findings highlight the need for standardized criteria and sleep measures to allow for more meaningful comparisons across studies and more powerful meta-analytic reviews.

It is the general consensus in the literature that SWDs can be considered primary to TBI and are not solely attributable to secondary physical or psychological symptoms (Beetar et al., 1996; Clinchot, Bogner, Mysiw, Fugate, & Corrigan, 1998; Fichtenberg, Millis, Mann, Zafonte, & Millard, 2000; Ouellet, Beaulieu-Bonneau, & Morin, 2006; Ponsford et al., 2012; Tham et al., 2012). The most comprehensive of these studies was conducted in 2007 by Baumann and colleagues, who investigated the risk factors and biomarkers for posttraumatic SWDs at 6 months post injury by obtaining neuropsychiatric and lab data, computerized tomography, sleep latency tests, PSG, and ACT from 65 patients with TBI at 4 days and 6 months post injury (Baumann et al., 2007). New onset SWDs were found in 72% of the patients and 60% of those patients had a SWD that could not be attributed to any specific cause (e.g., restless leg syndrome, OSA, behaviorally induced insufficiency sleep syndrome, depression) other than the TBI itself (Baumann et al., 2007). The cumulative impact of multiple brain injuries may also increase the severity of sleep disturbances. For example, Bryan (2013) reported that, compared to a single TBI and no TBI, multiple TBIs were associated with an increased risk for and severity of sleep difficulties in male service members in Iraq. Together, these findings suggest the effect of TBI on SWDs is cumulative and potentially causal.

Mild TBI patients, who comprise 70–90% of the treated TBI population (Cassidy et al., 2004), have been shown to have a higher rate of SWDs than those patients with more severe head injuries (Mahmood, Rapport, Hanks, & Fichtenberg, 2004; Orff et al., 2009; Ouellet et al., 2006; Parcell, Ponsford, Rajaratnam, & Redman, 2006). While TBI patients are generally susceptible to SWDs regardless of the severity of the head injury, one longitudinal study suggested that those with mild injuries are more likely to report symptoms of insomnia whereas those with more severe injuries are more susceptible to hypersomnia (Masson et al., 1996). Parcell, Ponsford, Rajaratnam, & Redman (2006) collected sleep diary data and found that individuals with TBI reported more nighttime awakenings, longer sleep onset latency, lighter (fragmented) sleep, and a change in bedtime (either earlier or later) compared to their pre-injury levels. This study also reported that those patients with milder brain injuries, amongst a mixed severity control sample, reported more nighttime awakenings. While some studies (Mahmood et al., 2004; Parcell et al., 2006) have found subjective sleep quality to be more severely impaired in mTBI compared to severe TBI, a recent study by Ponsford and colleagues (2013) did not replicate these findings. The authors suggest their findings might be explained by the fact that their mTBI group did not have higher anxiety or depression compared to the moderate and severe TBI group, suggesting the role of psychiatric symptoms contributing to increased sleep disturbance in mTBI. It is important to note that inconsistencies in the literature regarding the prevalence and severity of sleep disturbances across the TBI spectra, as in PTSD, may also reflect sleep state misperceptions (for a review see Orff et al., 2009).

In mTBI there is also evidence, albeit inconclusive, of circadian rhythm sleep disorders (CRSDs) (for a review see Orff et al., 2009). Ayalon and colleagues (2007) utilized PSG, ACT, temperature and melatonin data, as well as morningness/eveningness to measure the prevalence of CRSDs in 42 mTBI patients with complaints of insomnia. The authors reported a significantly higher incidence of CRSDs in the mTBI sample when compared to the prevalence amongst people attending sleep clinics with complaints of insomnia (36% vs 10%). Based on ACT data, 8 of the patients exhibited delayed sleep phase syndrome and 7 exhibited an irregular sleep-wake pattern. True to the clinical definitions, when compared to normative sleep-wake times, the data showed that the delayed sleep phase patients experienced delayed sleep onset (Mean = 2:33 am, Standard deviation = 76 mins) and offset times (Mean=11:57 am, Standard deviation =127 mins) while irregular sleep-wake pattern patients classically exhibited a disorganized and variable sleep and wakefulness episodes (day-to-day onset and offset variability > 2 hours). The incidences of delayed and irregular sleep phase were nearly equal in the mTBI sample, whereas amongst CRSD patients in sleep clinics rates are typically 85% and 2%, respectively (Dagan & Eisenstein, 1999). However, the authors note that patients with advanced or delayed sleep phase can exhibit normal sleep durations if allowed to choose their sleep schedule, which might mean that circadian shift disorders can go undetected if only sleep duration-related variables are examined out of the context of sleep pattern. While Ayalon and colleagues (2007) detected a significantly higher incidence of CRSDs and distinct temperature and melatonin profiles amongst the mTBI sample, Steele, Rajaratnam, Redman, and Ponsford (2005) reported no difference in sleep timing or melatonin onset in a sample of 10 TBI patients when compared to age and gender matched controls. However, the discrepant results between these two studies might be due to the fact that the majority of those from the latter study presented with moderate TBI. Moreover, anecdotally, one earlier case study published by Quinto, Gellido, Chokroverty and Masdeu (2000) describes the presence of new-onset delayed sleep phase syndrome at 4 years post mTBI. Despite a limited number of studies, these findings highlight the possibility of CRSDs following TBI.

While many TBI patients experience sleep disturbance, others report wake-related complaints such as fatigue, excessive daytime sleepiness, and hypersomnia (Baumann et al., 2007). Schreiber and colleagues (2008) reported increased daytime tiredness and a tendency to fall asleep during the day in a group of 26 mTBI patients compared to healthy matched controls. Kempf and colleagues (2010) followed a cohort of mixed severity TBI patients and reported that 67% of the patients who had endorsed sleep disturbances at 6 month post-injury continued to experience excessive daytime sleepiness, hypersomnia, fatigue, and decreased vigilance at 3 years post-injury. In fact, fatigue and hypersomnia increased at 3 years post injury and persisted up to 5 years post-injury (Bushnik, Englander, & Wright, 2008; Kempf et al., 2010). It has been reported that 20–33% of mTBI patients reported increased levels of fatigue compared to pre-injury levels (Englander, Hall, Stimpson, & Chaffin, 1992; Middleboe, Andersen, Birket-Smith, & Friis, 1992; Stulemeijer et al., 2006). It is not clear if wake-related complaints are the result of nocturnal sleep disturbance or a direct result of neuronal injury.

Recent works suggests biomarkers, such as relative cerebral metabolic rate of glucose, are potentially sensitive measures of sleep and wake anomalies in TBI. Stocker and colleagues (2014) used [¹⁸F]-fluorodeoxyglucose positron emission tomography to compare glucose metabolism during wakefulness, REM sleep, and NREM sleep in Veterans with a history of blast exposure without current concussive symptoms, relative to Veterans without a history of blast exposure. After controlling for PTSD symptoms, results revealed hypometabolism in limbic regions and visual cortices during wake and REM sleep in blast-exposed Veterans, despite similar PSG profiles and PSQI scores as the non-blast exposed group. Such findings suggest that there might be metabolic dysregulation during REM sleep and wake indicative of functional impairments in chronic blast-induced TBI—beyond stress or sleep symptoms—that would otherwise go undetected using standard sleep measures.

Conclusion

There are numerous possible refractory and protracted sleep disturbances in TBI, including hyposomnia, hypersomnia, and circadian dysrhythmia. Prior work has established a direct association between disturbed sleep and TBI beyond co-occurring psychological symptoms (Baumann et al., 2007). The mTBI population, which accounts for the majority of the TBI population, might be most affected by the sleep disturbances, compared to moderate and severe TBI. It is unclear how sleep disturbance severity interacts with the type and severity of brain injury as well as time post-injury. Further, the effects of head injury on sleep disturbance are cumulative, which is particularly problematic for returning Veterans with repetitive blast-induced mTBI. However, as in the PTSD literature, the inconsistent reports of the prevalence of SWDs in TBI might arise from sleep discrepancies between objective and subjectively measured sleep in TBI, suggesting sleep state misperceptions. More research is necessary to understand how objective and subjective measures can best characterize the incidence and persistence of SWDs across the spectrum of TBI severity. Further longitudinal studies and meta-analytic reviews with standardized sleep measurement criterion are also needed to delineate the complex temporal relationship between TBI and SWDs.

Pathophysiology of mTBI and relation to sleep

The prevalence of both sleep and arousal related complaints following TBI appear to be directly attributable to damage to sleep-regulation centers and arousal networks in the brain. Although these networks are mutually inhibitory, they are nonetheless distinct and it is important to investigate sleep and arousal related dysfunction both together and separately in the context of brain injury. In order to parse apart the factors underlying the complex relationship between sleep and TBI we briefly review the pathophysiology of mTBI in relation to sleep and arousal networks.

There are two types of traumatic brain damage; primary and secondary. Primary traumatic brain injury occurs at the moment of physical trauma and refers to the immediate mechanical load on the brain during rapid acceleration, deceleration, direct impact, penetration, and/or blast wave (Maas, Stocchetti, & Bullock, 2008). Primary damage affects the brain on both the tissue and cellular levels. Diffuse axonal injury, the most common form of primary tissue injury, results from rapid rotational acceleration of the brain that causes dynamic shearing of white matter tracts as well as tensile and compression strain (Johnson, Stewart, & Smith, 2013; Kan, Ling, & Lu, 2012). Other types of primary damage can include focal contusions, hematomas, and edema (Kan et al., 2012). Diffuse axonal injury is also of particular interest as it is common across the TBI severity spectrum (Johnson et al., 2013).

Following the impact of the trauma, primary injures can induce a cascade of secondary complications, which are associated with a myriad of biochemical processes including excitotoxicity, inflammation, free radical generation, oxidative stress, lipid peroxidation, hyperglycolysis, hyperglycemia, gene activation, neurotransmitter release, and calcium mediated damage (Johnson et al., 2013; Kan et al., 2012). Secondary damage can induce dysfunction at the mitochondrial level (Johnson et al., 2013; Kan et al., 2012; Maas et al., 2008). While the frontal lobes are the most susceptible to TBI-related injury (Dombovy, 2011; Marquez de la Plata et al., 2011), cortico-spinal strain and shear can inflict injury upon more distant, subcortical brain regions (Baumann, 2012). One of the most commonly affected areas following a TBI is the corpus callosum (Morey et al., 2013). Some have speculated that damage to the callosal fibers, which may aid in synchronization of interhemispheric connectivity during REM and NREM sleep (Bertini et al., 2004; Buchmann et al., 2011; Nielsen, Montplaisir, & Lassonde, 1992) and regulate the modulation of timing of REM sleep (Nielsen et al., 1992), contributes to disrupted sleep-wake processes.

Sleep promoting regions of the brain are still under investigation, however, the preoptic area including the ventrolateral (VLPO) and median (MNPO) preoptic areas have been shown to contain sleep-promoting neurons (Gong et al., 2004; D. McGinty et al., 2004; Sherin, Shiromani, McCarley, & Saper, 1996). The neurons in these regions are active during (NREM) sleep and are inactive during wakefulness (Suntsova, Szymusiak, Alam, Guzman-Marin, & McGinty, 2002; Szymusiak, Alam, Steininger, & McGinty, 1998; Takahashi, Lin, & Sakai, 2009). Lesion studies of the VLPO and MNPO have demonstrated a marked reduction in overall quantity and quality of sleep (Lu, Greco, Shiromani, & Saper, 2000; D. J. McGinty & Sterman, 1968). The VLPO and MNPO are thought to promote sleep by inhibiting arousal regions during sleep. Their nuclei release the inhibitory neurotransmitter gamma-aminobutyruc acid and the inhibitory neuropeptide galanin and innervate brain areas that function in arousal including the locus coeruleus, dorsal raphe, and the hypothalamic orexin/hypocretin neurons (Gaus, Strecker, Tate, Parker, & Saper, 2002; Saper, Fuller, Pedersen, Lu, & Scammell, 2010; Saper, Scammell, & Lu, 2005; Sherin, Elmquist, Torrealba, & Saper, 1998). This inhibitory interaction functions similarly to a flip-flop switch (Saper et al., 2001), by which a transition from wakefulness to sleep and vice versa is produced when the system is pushed into one direction or the other. However with neuronal loss or damage on either side of the switch, the result may be less stability in the inhibitory mechanisms of the switch, thus resulting in more transitions between sleep and wakefulness (i.e., frequent awakenings) (Chou et al., 2003). For example, some have suggested that age-related cell loss in the VLPO results in sleep fragmentation and increased daytime sleepiness due to weakening of the flip-flop switch (Saper, Cano, & Scammell, 2005).

It has been postulated that fatigue-related symptoms following TBI might have specific neurobiological underpinnings (Belmont, Agar, Hugeron, Gallais, & Azouvi, 2006). The wake promoting neurotransmitter hypocretin (orexin) is produced by a distinct set of neurons in the posterior hypothalamus (Baumann, 2012) and plays a crucial role in sustaining arousal, and preventing flip-flop between sleep-wake transition states (Saper, Scammell, et al., 2005) and fatigue-related symptoms following TBI may be due to damage to the system sustaining arousal via the “flip-flop” switch (Belmont et al., 2006). Low cerebrospinal fluid (CSF) hypocretin (orexin) has been reported in patients following TBI (Baumann et al., 2009; Baumann et al., 2005). While hypocretin (orexin) levels recovered in some patients at 6 months post-TBI, some patients showed only a partial recovery. Moreover, the reported low levels of CSF hypocretin at 6 months post injury were associated with excessive daytime sleepiness (Baumann et al., 2007).

As mentioned earlier, a growing body of evidence suggests dysregulation of BDNF in TBI and PTSD (for a review see Kaplan et al., 2010). There is evidence that upregulation of BDNF reduces the impact of secondary injury following TBI (Kaplan et al., 2010). However, there might be decreases in BDNF with increased time since injury, resulting from tissue loss over the first year post-injury (Schober, Block, Requena, Hale, & Lane, 2012). Previous work suggests a causal link between BDNF expression during wakefulness and subsequent slow-wave activity (SWA) during sleep (Faraguna et al., 2008). SWA is a sensitive marker for sleep need and NREM sleep intensity, as evidenced by increased SWA following prolonged daytime activity (Faraguna et al., 2008). Further, exogenous application of BDNF to one hemisphere in awake rats is associated with increased SWA during NREM in sleep the BDNF-injected hemisphere relative to the contralateral hemisphere and vehicle injection group (Faraguna et al., 2008). Given the foregoing findings, it is possible that downregulation of BDNF in chronic TBI or PTSD (or an interaction of the two) could decrease NREM sleep intensity and/or drive for sleep initiation.

Conclusion

Despite a multitude of secondary complications associated with TBI, it appears that SWDs may be linked to the neuronal injury itself, while psychiatric co-morbidities such as depression may serve as additional risk factors for post-traumatic SWDs (Baumann et al., 2007; Kempf et al., 2010). Incongruent findings between studies and wide ranges of SWD incidences reported in the literature may be due to heterogeneous methodologies including variability of: measures and methods used to obtain the data, sample size, and proportion of TBI severity groups within mixed samples (Baumann et al., 2007). Inconsistent findings and reported discrepancies in prevalence of mild and severe TBI, particularly by studies using subjective (self-report) sleep measures, may also be due to varying levels of anxiety, depression, or patient’s self-insight, as patients with mTBI may have increased awareness of post-injury health complications and may be more likely to report sleep disturbance as an injury-related problem (Tham et al., 2012). Moreover, studies that collapse data from the mixed severity samples may lack sensitivity to detect the differences in sleep disturbance presentation between mild, moderate, and severe TBI. Finally, mixed findings of sleep disturbances in TBI may reflect the differential neuronal injury profiles resulting from variable sources of insult (i.e., blast, motor vehicle accident, etc). Overall, research suggests that sleep and arousal deficits are clearly present in individuals with TBI regardless of severity, and there are likely neurobiological underpinnings to these difficulties that need to be investigated further in order to elucidate the mechanism of recovery and to improve outcomes in TBI.

Impact of sleep on cognition in TBI

In individuals with TBI, cognitive deficits are among the most frequently reported symptoms regardless of severity of the injury (Senathi-Raja, Ponsford, & Schonberger, 2010; Heidi Terrio, Nelson, Betthauser, Harwood, & Brenner, 2011). In mTBI, decrements in performance are found in processing speed, working memory, attention, memory, and executive functioning immediately post-injury (Frencham, Fox, & Maybery, 2005), but problems often decrease with time after injury (Brenner et al., 2010). Despite the evidence that performance on cognitive measures improves over time after mTBI, subjective complaints remain (Senathi-Raja et al., 2010; Heidi Terrio et al., 2009) raising the possibility that traditionally used objective measures may not be able to capture the subtle difficulties experienced by individuals with mTBI (Brenner, 2011). Additionally, it has been postulated that although cognitive performance among individuals with mTBI may not be substantially different than those without TBI, they may require more functional resources to complete tasks, such as requiring larger areas of their cortex for task completion (Van Boven et al., 2009).

Sleep difficulties in those with mTBI may further impact their cognitive functioning and may contribute to the variability in cognitive performance, which may also explain the reported discrepancies in the literature discussed earlier in this review. Individuals with TBI are particularly vulnerable to sleep difficulties, and some evidence suggests that those with mTBI may be at even greater risk for sleep disturbances compared to more severe TBI, due to greater self-insight and therefore may present with increased vulnerability to sleep deficits associated with psychological and emotional factors (Bloomfield, Espie, & Evans, 2010). Indeed, at least one study examining whether sleep difficulties exacerbate deficits in sustained attention following TBI found that when individuals with TBI were separated into good and poor sleepers, based on both subjective and objective measures of sleep, poor sleepers performed significantly worse on measures of sustained attention (Bloomfield et al., 2010). A large proportion of individuals coming from the OEF/OIF/OND combat arenas present with both PTSD and mTBI and approximately 40% of PTSD-mTBI Veterans report severe sleep disturbances (Wallace et al., 2011) as well as cognitive problems (Schiehser et al., 2011; Spencer, Drag, Walker, & Bieliauskas, 2010). Therefore, the examination of the relation between their cognitive difficulties and sleep problems is particularly important. Kark and colleagues (2013) examined the relationship between sleep, PTSD symptom severity, daily cognition, and quality of life in OEF/OIF/OND Veterans with exposure to blast and current cognitive complaints. Veterans with PTSD had worse sleep on both subjective and objective measures of sleep quality and poor sleep was related to worse self-reported daily cognition, poorer community integration, and reduced satisfaction with life.

Conclusion

Cognitive deficits, namely attention, concentration, and executive dysfunction, are the primary complaints of individuals with TBI. There is evidence that these deficits may be exacerbated by poor sleep, however few studies have examined the association between poor sleep following TBI and cognition. It is also unclear how severity, type, and number of brain injuries might contribute to objective and subjective cognitive impairments differently. Future work is needed to not only understand how poor sleep perpetuates poor cognition, but also how cognitive impairment might effect sleep (e.g., elevated stress related to adverse functional consequences of executive dysfunction).

Sleep interventions in TBI

Studies of pharmacologic treatment in TBI have mostly focused on treatment of fatigue and daytime somnolence associated with moderate to severe TBI. One study examined individualized treatment of patients with TBI that involved either pharmacologic or behavioral treatment and found that effective treatment optimized recovery and outcomes for adults with chronic TBI. Notably, the authors did not specify the modality of treatments used (Wiseman-Hakes et al., 2013).

There is a paucity of studies examining the efficacy of treatments aimed at alleviating sleep difficulties and fatigue in individuals with TBI. One single-case experimental design study found that CBT-I reduced nocturnal sleep difficulties in one individual with TBI (Ouellet & Morin, 2007). A recent study examining the impact of sleep interventions in a small group of moderate to severe TBI (n=10) and 2 mTBI patients, found that individualized sleep interventions that included sleep hygiene recommendations, pharmacological interventions and/or treatments for sleep apnea, was associated with improvements in insomnia severity, depression symptoms, speed of cognitive processing, and language functions (Wiseman-Hakes et al., 2013). As with other individuals with sleep difficulties, long term use of pharmacologic agents for sleep is not generally recommended and is particularly not recommended for individuals with TBI due to side effects such as impaired daytime alertness and cognitive function that could prove an additional burden (Flanagan, Greenwald, & Wieber, 2007; Larson & Zollman, 2010). Further careful consideration is required for treatment of sleep in comorbid TBI and PTSD, since some traditionally prescribed PTSD medications may further exacerbate the TBI- and/or treatment-related cognitive problems. One study examined the effect of melatonin on sleep and daytime alertness in comparison to amitriptyline. This study found no improvement in sleep quality with either treatment, but found that melatonin had a greater effect on daytime alertness than amitriptyline (Kemp, Biswas, Neumann, & Coughlan, 2004). Another study recently reported that relaxation treatment in OEF/OIF/OND Veterans with mTBI was an acceptable milieu by both patient and provider, although the Veterans preferred a combination of relaxation and pharmacologic treatment (Epstein, Babcock-Parziale, Haynes, & Herb, 2012).

There is emerging evidence for the therapeutic utility of novel non-invasive neuromodulation interventions, such as low-level laser therapy (LLLT) in sleep disorders. Therapeutic effects of LLLT using red light-emitting diodes (LED) on sleep were demonstrated in healthy adults (Zhao, Tian, Nie, Xu, & Liu, 2012) and patients with insomnia (Xu , Wang, Ta, & Li, 2001; Xu, Wu, Wang, Shang, & Li, 2002). We are aware of one pilot study reporting improved sleep and PTSD symptoms following the red/near infra-red LED treatment in patients with chronic moderate TBI and comorbid PTSD (Bogdanova et al., 2014). The study results suggest that sleep structure (ACT recordings) can be changed by transcranial LED treatment in patients with chronic TBI, as sleep efficiency increased one week post-LED treatment and was maintained at 2-month follow-up. LED is a non-invasive, portable and relatively inexpensive neuromodulation treatment with potential to reduce multi-symptoms in patients with TBI and associated comorbidities, such as sleep disturbance and PTSD. More research and randomized controlled studies are needed to investigate specific LED effects on sleep and other neuropsychiatric symptoms in this population.

Treatment with bright light to improve daytime function and alertness has been suggested in TBI, but there were no published studies evidencing its efficacy (Ponsford et al., 2012). A recent study evaluated efficacy of short wavelength (blue) light therapy on fatigue in patients with TBI who self-reported fatigue and/or sleep disturbance (Sinclair, Ponsford, Taffe, Lockley, & Rajaratnam, 2014). Researchers reported that the high-intensity blue light therapy resulted in reduced fatigue and daytime sleepiness during the treatment phase, with a trend toward baseline levels 4 weeks after the treatment. The authors concluded that the light therapy appears to be effective in alleviating fatigue and daytime sleepiness following TBI.

Taken together these and other preliminary findings suggest that novel photobiomodulation therapeutic approaches have the potential to reduce sleep disturbance in patients with TBI and associated neuropsychiatric symptoms.

Conclusion

Sleep disturbance presents significant problems for the returning Veterans with TBI and associated neuropsychiatric comorbidities, such as PTSD and depression. Given that sleep plays an important role in the physiologic processes underlying neural recovery (Parcell, Ponsford, Redman, & Rajaratnam, 2008), it is critical to identify and treat sleep disorders in patients with TBI early in order to promote and facilitate their recovery. As we discussed earlier, pharmacologic management has been the main treatment option for sleep disorders for patients with insomnia and other neurologic and neuropsychiatric disorders. There is limited evidence for treatments specifically aimed at alleviating sleep disturbance in TBI or TBI with PTSD. Recent studies suggest that novel neuromodulation treatments, such as LED or light therapy, can improve sleep in TBI, and may potentially provide a safe and non-invasive non-pharmacological treatment for patients with TBI and associated neuropsychiatric comorbidities.

CHALLENGES AND FUTURE DIRECTIONS

There is currently a need for research to identify best practices for treatment of patients with mTBI and associated comorbidities, such as PTSD and sleep disturbance (Report of VA Consensus Conference, 2010). An interdisciplinary team approach can provide the necessary resources for diagnosis and treatment of multiple comorbidities in this population (Institute of Medicine, 2011). Future efforts should focus on the development of new treatment protocols and multi-modal treatment programs that could be individually tailored to better serve the patient’s individual needs.

One of the challenges in treatment planning for returning Veterans with TBI, PTSD, and sleep difficulties is determining the optimal timing for treatment of each of the multiple health issues. The argument in favor of early treatment of cognitive dysfunction may be based on the premise that TBI–related cognitive deficits may diminish treatment response to PTSD interventions, such as cognitive-behavioral treatment or exposure therapy (Bogdanova & Verfaellie, 2012). Alternatively, PTSD treatment may become a priority for patients with more severe PTSD symptomatology, since problems with impulse control and emotion regulation may affect patient’s ability to engage in cognitive treatment (Bryant & Hopwood, 2006). Moreover, there is a possibility that PTSD interventions may also help to improve sleep but in most cases sleep problems remain at clinical levels (Belleville, Guay, & Marchand, 2011; Galovski, Monson, Bruce, & Resick, 2009; Zayfert & DeViva, 2004) even after successful PTSD treatment. However, if sleep problems do indeed stand in the way of successful recovery from PTSD and mTBI, then sleep may become the principal target of treatment, possibly preceding the start of intervention targeting the primary symptoms for each condition. Future clinical trials may provide a much needed evidence base.

Another key issue to consider is treatment accessibility. Providing treatment to patients in rural locations might be possible utilizing web-based treatment paradigms. Though limited, there is emerging evidence that web-based CBT with automated support may be effective in improving the sleep and associated daytime functioning of adults with insomnia (Espie et al., 2012). Tele-health protocols specifically targeting sleep difficulties in Veterans with multiple comorbidities should be developed and implemented. Three recent research reports demonstrate that up to eight weekly sessions of telephone-delivered CBT-I (Arnedt et al., 2013) or five weeks of computerized CBT-I (Vincent & Walsh, 2013) or six weeks of unsupported, internet-delivered self-help CBT-I (Lancee, van den Bout, van Straten, & Spoormaker, 2012) can improve sleep in adults with chronic insomnia. It would be also important to evaluate whether such protocols can be utilized in treating active duty military personnel. For example, one recent study reported efficacy of a behavioral intervention for insomnia in OEF/OIF/OND Veterans that involved one in-person and three telephone sessions as well as electronic delivery components (Epstein et al., 2013). The authors reported large effect sizes for pre- to post-treatment decrease in insomnia severity and sleep improvements were maintained at 3-months post-treatment.

Further research is needed to develop treatment programs that would optimize outcome and increase adherence to the treatment, which presents significant challenge for Veterans with mTBI/PTSD and their providers (Sayer et al., 2009). PTSD symptoms, such as emotional dysregulation and avoidance may significantly limit and disrupt the treatment of sleep or cognitive problems. Web-based treatment programs, as well as integrated programs implemented in primary care clinical settings, might be easier for Veterans to handle. Randomized clinical trials are needed to evaluate efficacy of such programs in Veteran population.

While there is currently much interest in potential biomarkers for the diagnosis and prognosis of outcome in TBI, this area is still under development. Research is needed to develop methods for identification of predictors and moderators of natural recovery and treatment outcomes in Veterans with TBI and multiple comorbidities, such as sleep problems and PTSD, so that treatment programs can be optimized to fit the patient’s needs. Future research efforts can elucidate the neural bases of sleep interventions and the mechanisms of change underlying the efficacy of sleep treatments, and inform the future treatment programs development.

SUMMARY

TBI and PTSD are currently in the spotlight as problems impacting the current wave of Veterans returning from the OEF/OIF/OND combat arenas. However, neither condition is specific to this group of Veterans as both have been endemic to combat and non-combat Veterans through generations. More recently, sleep has gained the attention of researchers and clinicians from a variety of disciplines, being recognized as bi-directionally related to multiple psychological and medical conditions, as well as mortality. While sleep problems have been shown to impact quality of life and functional status on many levels, including psychological, cognitive, and physical functioning, it is also understood that sleep can be disrupted by multiple physical and psychological conditions.

Sleep is among the primary symptoms of PTSD and TBI, and one of the mediating factors affecting the level of severity and treatment outcome of both conditions. Despite its high prevalence, few treatments and rehabilitation protocols target sleep as a primary symptom; although in many cases even after the resolution of mTBI or PTSD, poor sleep remains at clinical levels (Belanger, Morin, Langlois, & Ladouceur, 2004; Clum, Nishith, & Resick, 2001; Kramer, Booth, Han, & Williams, 2003; Wittmann, Schredl, & Kramer, 2007) and continues to impact daily functioning and quality of life (Belanger, Morin, Langlois, & Ladouceur, 2004; Clum, Nishith, & Resick, 2001; Kramer, Booth, Han, & Williams, 2003; Wittmann, Schredl, & Kramer, 2007).

One of the key problems in sleep research is that sleep is not a stable state, and sleep need and drive vary significantly between individuals, especially those with neurologic and neuropsychiatric disorders. As a result, the research on sleep, which is limited by the nature of the sleep state and sleep measurement (i.e., objective vs. subjective reporting; inter-individual and within-individual differences), has resulted in many inconsistencies among studies and a lack of consensus on a variety of sleep-related issues, including treatment recommendations. Future research efforts should target the development of sleep-focused interventions and multi-modal treatment programs adjustable to the specific needs of patients with multiple comorbidities, such as TBI, PTSD, depression, and sleep disturbance.

Research suggests that sleep is necessary for the consolidation of memories and is important for cognitive, psychological, and physical functioning, which leads to the main point presented in this review: evaluation and treatment of sleep problems in TBI and PTSD should become an integral part of clinical management of these disorders. Although the treatment of sleep problems has long been rooted in pharmacology, more recently developed non-pharmacologic interventions have proven effective in treating sleep, and may become efficient clinical tools to improve daily functioning, treatment outcome, and quality of life in individuals with TBI and PTSD.

Highlights.

• Poor sleep may impact recovery and treatment outcome in Veterans with PTSD and TBI

• Sleep difficulties are a risk factor for the development and maintenance of PTSD

• TBI patients can suffer permanent sleep problems regardless of injury severity

• Sleep-focused interventions could augment current TBI and PTSD treatment protocols

• We review physiological mechanisms and treatment of sleep problems in TBI and PTSD

Abbreviations

ACTH

Adrenocorticotropic hormone

BDNF

Brain-derived neurotrophic factor

CRSD

Circadian rhythm sleep disorders

CBT-I

Cognitive Behavioral Therapy for Insomnia

CAR

Cortisol awakening response

CSF

Cerebrospinal Fluid

HPA

Hypothalamic-Pituitary-Adrenal axis

IRT

Imagery Rehearsal Therapy

LED

Light emitting diodes

LLLT

Low-level laser therapy

MNPO

Median preoptic area

OSA

Obstructive sleep apnea

OEF/OIF/OND

Operations Enduring Freedom and Iraqi Freedom and Operation New Dawn

PSG

Polysomnography

PTSD

Post-Traumatic Stress Disorder

REM

Rapid eye movement

NREM

Non-rapid eye movement

SWDs

Sleep-wake disturbances

SWA

Slow wave activity

SWS

Slow wave sleep

TBI

Traumatic brain injury

TST

Total sleep time

VLPO

Ventrolateral preoptic area