© Springer Science+Business Media LLC 2018Eric Vermetten, Anne Germain and Thomas C. Neylan (eds.)Sleep and Combat-Related Post Traumatic Stress Disorderdoi.org/10.1007/978-1-4939-7148-0_15
15. Sleep Disturbances and Sleep Assessment Methods in PTSD
Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA
University of Pittsburgh Medical Center, Pittsburgh, PA, USA
KeywordsSleep disturbancesAssessment methodsPTSDSleep duration
Sleep is a universal, biologically driven, and multifaceted state of behavioral quiescence that supports mental and physiological health and well-being. A detailed summary of the multiplicity of neurobiological events that are required to initiate and sustain consolidated and restorative sleep exceeds the scope of this chapter (see [1, 2] for review). Different observable dimensions of sleep, however, are direct expressions of the complex, recurrent, and reversible neurobiological processes that support the experience of sleep. Dimensions such as sleep duration, consolidation, timing, and sleep quality have been proposed as related yet distinct dimensions that contribute to the overall experience of sleep . Healthy sleep, thus, can be defined by regular, predictable, sufficient, and restorative or non-fragmented sleep episodes. All of these dimensions are disrupted in combat-exposed service members and veterans with posttraumatic stress disorder (PTSD).
Self-report and objective sleep measurement methods capture different aspects of these dimensions of sleep. However, none of the available sleep measurement methods can simultaneously capture all dimensions of sleep with high concordance between subjective and objective methods. The discordance between measurement methods is even greater in clinical samples . Unfortunately, this discordance has been misinterpreted as a “misperception” of sleep in individuals with PTSD . Specifically, subjective reports of sleep disturbances are considered to be inaccurate (i.e., “misperceived”) as few objective polysomnographic (PSG) anomalies have been consistently found in adults with PTSD, and none are specific to the disorder [6, 7]. However, sleep neuroimaging studies have shown that PSG does not capture subcortical metabolic and blood flow changes that distinguish between healthy and clinical samples with poor sleep. Cortical and subcortical differences between groups of healthy sleepers and groups of individuals with psychiatric or sleep disorders can be detected even when PSG measures fail to distinguish between healthy and clinical groups [8–12]. Thus, the inconsistencies between subjective reports and objective PSG more likely reflect the limitations of PSG, rather than the absence of meaningful changes in brain activity during sleep in individuals with PTSD.
The measurement of sleep and sleep disturbances in PTSD, thus, requires careful considerations of the specific objectives of assessment. For clinical purposes, self-report clinician-administered sleep assessments and clinical sleep evaluations provide sufficient information to accurately define and diagnose insomnia, nightmares, and other parasomnias that are common among trauma-exposed individuals. Other sleep disorders, such as sleep-disordered breathing (SDB), periodic limb movement disorder (PLMD ), narcolepsy, rapid eye movement (REM) sleep behavior disorder (RBD), or trauma-associated sleep disorder (TASD), can be diagnosed based on clinical and self-report measures, but require confirmation with specific in-lab polysomnographic studies. For research questions that target sleep-specific mechanisms and correlates of PTSD, more resource-intensive measurement methods such as actigraphy, PSG, or sleep neuroimaging methods must be considered.
In this chapter, the types of sleep disturbances observed in adults with PTSD are first briefly described. Advantages and limitations of assessing sleep disturbances using items extracted from PTSD-specific measures are then discussed. Sleep-specific measures that provide a more detailed evaluation of the nature and severity of sleep disturbances in PTSD and that align more closely with the diagnostic features of sleep disorders are then provided. In this section, self-report instruments, prospective sleep diaries and actigraphy, and objective sleep measurement methods are then described. We conclude by offering specific recommendations for the assessment of sleep disturbances in service members and veterans with PTSD that align with clinical and research purposes.
Sleep Disturbances in PTSD
Insomnia and nightmares are two well-recognized core symptoms of PTSD. Both insomnia and distressing dreams are more prevalent in military samples relative to prevalence estimates reported in the general civilian population (e.g., [13–15]). Furthermore, both sleep disorders can first occur during deployment (see Chaps. 6 and 7) and prior to the onset of PTSD. In fact, both insomnia and nightmares increase the risk of developing PTSD following trauma exposure . For example, Gehrman and colleagues showed that service members with insomnia, short sleep duration, or both prior to deployment were at greater risk for developing PTSD post-deployment compared to their peers without insomnia or short sleep duration. These relationships persisted above and beyond the impact of combat exposure during deployment . In another study, van Liempt and colleagues showed that pre-deployment nightmares, but not insomnia, predicted PTSD symptom 6 months post-deployment . Following deployment, Sleep disturbances:insomnia and nightmares insomnia and nightmares also predict the severity of PTSD symptoms at subsequent assessments (e.g., [14, 15, 18]).
Other sleep disturbances common in PTSD but are often overlooked include SDB, sleep-related movement disorders, and other parasomnias. SDB refers to complete cessation of breathing (i.e., apnea) or marked limitations of airflow during sleep. Common risk factors associated with SDB include being male, older in age (i.e., >50 years old), obesity, and endorsing excessive daytime sleepiness. However, the performance of the first three of these indicators in military samples is questionable, as many younger, nonobese, and female service members and veterans are found to have clinically significant SDB. In the context of PTSD, daytime sleepiness due to sleep fragmentation is often masked by hyperarousal. Nevertheless, Colvonen et al. found that over 60% of veterans referred to a PTSD clinic screened positive for obstructive sleep apnea (OSA) and that one’s risk for OSA increased as a function of PTSD symptom severity . In another study, PTSD severity did not increase the risk for OSA, but over 70% participants seeking PTSD treatment were at risk for OSA .
PLMD has also been observed in veterans and civilians with PTSD [21–24]. This sleep-related movement disorder refers to the recurrent movements of the legs and/or arms at intervals of 20–40 s during non-REM (NREM) sleep . These movements are typically associated with brief arousals that are not remembered upon awakening. As a result, affected individuals report daytime fatigue, difficulty maintaining sleep, and non-restorative sleep.
Mysliewic and colleagues recently identified a new trauma-related REM sleep parasomnia called trauma-associated sleep disorder (TASD; Chap. 18) . Behaviorally, TASD resembles RBD by virtue of complex vocal and motor behaviors occurring during REM sleep that are associated with the loss of muscle atonia. However, TASD is specifically associated with trauma exposure and distinct polysomnographic and physiological features that accompany these episodes. TASD also does not appear to respond to first-line PTSD or RBD treatments , but prazosin may be effective.
Surprisingly, little is known about the prevalence of NREM sleep parasomnias, such as nocturnal panic attacks or somnambulism, in relation to PTSD in military samples. In civilian samples, a history of trauma is associated with increased prevalence of nocturnal panic attacks . Similarly, the prevalence of complex motor behaviors or vocalization during NREM sleep in trauma-exposed individuals with or without PTSD is unknown. One case report described 12 military trainees who were honorably discharged due to severe somnambulism . Although the literature on NREM sleep parasomnias is greatly limited, it is likely that trauma exposure and the sleep disruption and/or sleep curtailment associated with PTSD increase the risk of such episodes which, in turn, may be mistaken for nightmares. Similarly, it is likely that individuals with a personal and familial history of NREM sleep parasomnias would be especially at risk for the recurrence or exacerbation of these sleep disorders in the context of PTSD.
Assessing Sleep Disturbances Using Items Extracted from PTSD Measures
Over the past decade, there has been a growing recognition that sleep disturbances may not only reflect secondary symptoms or conditions in PTSD but may also reflect more critical aspects of the pathophysiology underlying this debilitating disorder and/or comorbid sleep disorders that require targeted clinical attention. This has encouraged many clinical investigators to extract sleep item scores from well-established PTSD symptom measures in order to cross-sectionally or prospectively explore the relationship between sleep, the severity of daytime PTSD symptoms, and clinical outcomes [13, 29–32].
This approach has the advantages of (1) limiting the number of additional assessments required to get a broad evaluation of sleep quality or severity of insomnia and nightmares, (2) aligning directly with other scales used to assess daytime symptoms of PTSD, (3) facilitating the collection of sleep data in large samples where additional or more time-consuming resources are typically impractical, and (4) providing a measure that captures improvements in sleep symptoms with treatments [e.g., 30, 32]. However, this economical approach does not provide well-validated information regarding different dimensions of sleep (i.e., quality, duration, timing, consolidation) nor does it allow for ruling out the presence of sleep disorders that may masquerade as insomnia or nightmares [22, 33].
The gold standard for assessing the presence and severity of PTSD is the Clinician-Administered PTSD Scale (CAPS) [34, 35]. The CAPS for the Diagnostic and Statistical Manual of Mental Disorders (DSM), 4th Edition, includes two sleep-related items. Item B-2 states “recurrent distressing or unpleasant dreams of the event,” where a frequency of at least once or twice for the period under assessment (i.e., past week, past month, or lifetime), is scored as “1” on this item. The intensity of the nightmare episodes is determined based on associated distress and difficulty returning to sleep. In the revised CAPS for DSM-5 , a moderate rating (i.e., clinically significant) requires the endorsement of at least two nightmares per month and the presence of distress and persistence of nightmare-related images or emotions during the day. The difficulty returning to sleep after awakening from unpleasant dreams is no longer specified as an index of nightmare severity.
Former item D-1 inquired about difficulty falling or staying asleep. Additional information was obtained to determine the extent to which difficulties initiating or maintaining sleep or early morning awakenings were present. The number of desired vs. actual hours of sleep was also assessed, presumably to evaluate sleep duration/loss. The intensity score for item D-1 was based on the estimated sleep loss, where a sleep loss of up to 30 min was rated as “mild” (score = 1) and a sleep loss of 90 min to 3 h was considered severe (score = 3). In the CAPS-5, the new sleep item E6 is similar to the former D-1 item in focusing on sleep latency, nocturnal awakening, early morning awakenings, and total sleep or estimated sleep loss. Interestingly, the header to this item now specifies “restless sleep,” but a definition is not provided.
While the use of the individual or combined sleep item scores from the CAPS has face validity, a significant problem inherent in this approach is that the anchors provided to determine clinically significant frequency (score = 1) and intensity (score = 2) do not align with the thresholds of clinical significance established for nightmare disorder or insomnia . For instance, a frequency of a least one nightmare per week with any sleep disruption of unspecified duration and at least one instance of sleep disruption, distress, or other daytime impairment would indicate clinically significant nightmares. Thus, the anchors provided with the CAPS may underestimate the severity and clinical relevance of nightmares experienced by individuals with PTSD.
A similar but inverted issue arises when estimating insomnia with item D-1 only. A score of “1” for frequency reflects insomnia that occurs less frequently than three times per week, which represents the threshold established by the International Classification of Sleep Disorders, 3rd edition, and DSM-5 [25, 36]. Furthermore, the cutoff regarding sleep loss provided for scoring the intensity of sleep disturbances is arbitrarily determined; the diagnostic and research criteria for insomnia do not require a range of duration of reported sleep loss. A sleep latency or wake time after the initial onset of sleep of >30 min is considered clinically significant . Furthermore, other indicators of the impact of sleep loss reported in insomnia, such as daytime impairments in occupational or social functioning and cognitive, somatic, or mood symptoms, are not considered in the rating of the intensity of the sleep disturbance. Thus, item D-1 provides an aggregate estimate of insomnia features and sleep loss that can be helpful, but is not a good estimate of insomnia as clinically defined. While the use of these CAPS items can provide a simple global estimate of sleep in PTSD, this approach does not accurately capture the phenomena of interest, nor does it provide sufficient information to detail sleep among different dimensions in PTSD to guide decisions for further sleep assessment of treatment planning. Thus, when used to assess the impacts of treatment of sleep in PTSD [30–32], the use of the CAPS sleep items (individually or in combination) may inaccurately estimate the magnitude of the severity of sleep disturbances as well as treatment-related improvements in sleep. More comprehensive measures should be considered especially in the context of clinical trials.
Extracting sleep items from self-report measures of PTSD symptom severity to evaluate sleep disturbances is also a common practice, and this approach faces similar limitations to extracting sleep items from the CAPS. Self-report PTSD measures with sleep items include teh PTSD Checklist [38, 39], the Davidson Trauma Scale , the Impact of Event Scale (IES) [41, 42], and the Mississippi Scale for Combat-Related PTSD (M-PTSD ) . Some of these instruments assess distress associated with sleep disturbances and unpleasant dreams, while others assess the frequency and severity of these symptoms [41, 42, 44]. The time frame used as reference to determine the frequency of symptoms varies significantly across instruments (past 2 weeks to past 6 months), which makes it difficult, at best, to determine an adequate sleep-focused intervention plan.
In summary, extracting sleep items from PTSD assessment instruments can provide a broad assessment of some aspects of sleep disturbances, but does not provide information that aligns with the diagnostic criteria for insomnia or nightmares, and only captures certain aspects of sleep-related daytime impairments. To better characterize sleep disturbances in PTSD, validated, sleep-focused instruments are recommended. Some of the instruments recommended for the assessment of insomnia, unpleasant dreams, and other sleep disturbances are detailed in the next section.
Self-Report Measures of Insomnia, Sleep Quality, and Disruptive Nocturnal Behaviors
A number of self-report measures of sleep and sleep disturbances have been validated in civilian samples and used in military samples. These tools are brief and capture global sleep quality through specific dimensions of sleep or severity of sleep disorders. They typically have cutoff scores to determine the clinical significance of sleep disturbances. It is important to note, however, that only a few studies have assessed the validity of cutoff scores for these instruments in military samples, and none has been validated in active duty military samples, where sleep is often curtailed or disrupted by occupational, operational, and/or environmental demands, and circumstances are common sleep challenges. Nevertheless, they offer a starting point to evaluate overall sleep patterns and disturbances. A handful of the recommended assessment tools and methods are summarized here, but the reader is encouraged to consult expert recommendations for the assessment and management of sleep disorders [45–48].
The Insomnia Severity Index (ISI) [49, 50] is the recommended self-report measure to capture the severity of insomnia. The ISI is a seven-item measure that captures both nighttime and daytime disturbances that characterize insomnia [25, 36]. Scores range from 0 (no insomnia) to 28 (severe insomnia). Scores from 0 to 7 indicate no symptoms of insomnia; 8–14, mild; 15–21, moderate; and scores above 21 severe insomnia. A reduction of eight points (or a decrease by one severity category) has been proposed as a clinically meaningful improvement in insomnia . In clinical trials, remission can be defined as an end-point ISI score of less than or equal to 7. The psychometric properties of the ISI have not been evaluated in military samples.
The Pittsburgh Sleep Quality Index ( PSQI ; ) is the most widely used self-report measure of global sleep quality. The PSQI is an 18-item self-report measure that broadly assesses global sleep quality and includes subscales to evaluate sleep duration, sleep onset latency, sleep efficiency (ratio of total sleep time/total time spent in bed), sleep disturbances, daytime impairments, and use of sleep medications. The time reference is 1 month, and each item and summary subscale are scored from 0 (not in the past month) to 3 (three or more times/week). Total scores range from 0 (good sleep quality) to 21 (poor sleep quality). A cutoff score of greater than or equal to 5 is associated with clinically significant sleep disturbances in civilians. In samples of civilians and military veterans with PTSD, mean PSQI scores typically range from 10 to 15 [29, 33, 53–55]. A reduction in PSQI scores by three points has been suggested as a minimally meaningful difference in sleep-focused clinical trials [56–58].
An addendum to the PSQI was specifically designed to capture trauma-related sleep disturbances and has been validated in civilian [59–61] and military samples [62, 63]. The PSQI Addendum for PTSD (PSQI-A) is a seven-item self-report measure that assesses the frequency of disruptive nocturnal behaviors , including nocturnal panic attacks , night terrors, acting out dreams, night sweats, nightmares or bad dreams related to traumatic events, unpleasant dreams unrelated to traumatic events, intrusive thoughts and images at night, and general nervousness. Like the PSQI, each item of the PSQI-A is rated on a three-point scale. The sum of item scores ranges from 0 (no disruptive nocturnal behaviors) to 21 (severe disruptive nocturnal behaviors). A cutoff score of 4 has shown high specificity and sensitivity to identify individuals with and without PTSD [59, 62]. The PSQI-A has thus been proposed as a brief, sleep-focused screening measure of PTSD. Although the PSQI-A is sensitive to change with sleep treatment, a minimally meaningful difference has not been established.
For a more focused assessment of nightmares and unpleasant dreams , the Disturbing Dream and Nightmare Severity Index ( DDNSI) measures the frequency, intensity, and severity of nightmares and unpleasant dreams. For each, an estimate of the number of dreams per night and total number of nights with nightmares/unpleasant dreams per week is assessed. The severity of these episodes is rated on a scale from 0 (no problem) to 6 (extremely problematic). The occurrence of awakening following unpleasant dreams is rated on a four-point scale (0 = never to 4 = always). An index score is obtained by adding the number of nightmares per week and the number of nights with nightmares per week, along with severity of nightmares and frequency of awakening. A score of 10 or greater indicates the presence of nightmare disorder [64, 65].
Other Self-Report Sleep Measures
Because sleep apnea is prevalent in military samples and can be mistaken for insomnia, nightmares, depression, or daytime fatigue, a brief screening measure can be helpful to detect SDB in this at-risk population. The STOP-BANG and Berlin Questionnaire offer well-validated instruments to assess the risk of sleep apnea. The STOP-BANG [66, 67] is a questionnaire that includes eight dichotomous items (yes/no) to assess the risk of clinically significant sleep apnea, based on snoring, tiredness or sleepiness, observed sleep apneas, high blood pressure, body mass index, neck circumference, age, and gender. A positive answer to three of the eight items suggests the presence of moderate risk of sleep apnea. The Berlin Questionnaire  is a ten-item questionnaire that assesses the presence and characteristics of snoring, daytime sleepiness, and hypertension. Positive scores in two of these three categories suggest a high risk of sleep apnea.
Questionnaires to preliminarily assess symptoms of sleep disorders are often lengthy (e.g., Sleep Disorders Questionnaire; ) and impractical in research or general clinical settings and offer little advantage over clinical interviews.
Prospective Sleep Assessments
Questionnaires , like clinical interviews, only provide a snapshot of sleep disturbances and different dimensions of sleep. Prospective measurement methods have the advantage of providing longitudinal information that captures both inter- and intraindividual variations over several sleep/wake cycles, sleep duration regularity, efficiency, and sleep quality over extended periods of time.
Sleep diaries are a critical clinical tool in the assessment and treatment of insomnia, nightmares, and other sleep disorders . Multiple versions of sleep diaries exist, and one has been created by consensus of experts  (see also ). Upon awakening in the morning, individuals are asked to provide information about the previous night’s sleep latency, total time in bed and total sleep duration, number and duration of nocturnal awakenings, sleep quality, and restfulness upon final awakening. Other events such as dreams and nightmares can also be logged. For a representative sampling of sleep/wake patterns, completing the diary for 7–14 days is recommended. Sleep diaries can also be augmented by an evening section that is completed prior to bedtime to monitor sleep-impacting behaviors, such as napping, caffeine and alcohol use, use of medications, and stress levels during the day. While a paper diary may not be completed in a timely fashion, several online or electronic diaries are now available and provide time stamps to ascertain the timeliness of entries. Sleep diaries are helpful to capture variations in sleep/wake patterns, provide a more fine-grained portrait of overall sleep habits, and track adherence to treatment recommendations and improvements over time.
On the other hand, actigraphy is an objective and prospective measure of rest-activity patterns. Actigraphs are wristwatch-like devices that use accelerometer to capture motion from the nondominant wrist. Rest-activity patterns are derived from periods of inactivity (used as a proxy for sleep episodes) and activity (used as a proxy for wakefulness).
Actigraphs can be worn for several consecutive 24-h periods, without the need to recharge or download data collected over time. Actigraphy has the advantage of providing a low-burden means to capture the regularity, timing, and duration of rest-activity cycles over time, but do not capture sleep quality. While actigraphy is of low burden for users, the burden on clinician and research resources is not negligible, given each device must be calibrated and have data downloaded and manually processed. Determining the proper acquisition parameters and scoring protocols for actigraphy is also crucial for obtaining reliable and replicable data  (e.g., [73, 74]). Concurrent acquisition of actigraphy with an event marker to signal the onset and offset of the rest period and completion of a sleep diary to identify the main rest period can significantly improve the concordance between the two measurement methods [73, 74] but can rapidly become resource intensive. In veterans with PTSD, poor correlations between actigraphy and sleep diary data have been reported .