Sleep disturbances as predictors of prolonged exposure therapy effectiveness among veterans with PTSD

Psychiatry Research 256 (2017) 118–123

Contents lists available at ScienceDirect

Psychiatry Research journal homepage: www.elsevier.com/locate/psychres

Sleep disturbances as predictors of prolonged exposure therapy effectiveness among veterans with PTSD

MARK

⁎

Minden B. Sextona,b, , Kimberly M. Avallonea, Erin R. Smitha,b, Katherine E. Portera,b, Lisham Ashrafiouna, J. Todd Arnedtb, Sheila A.M. Raucha,b,c,d a

Ann Arbor Veterans Healthcare System, Mental Health Service (116C), 2215 Fuller Rd., Ann Arbor, MI 48105, USA University of Michigan Medical School, Departments of Psychiatry and Neurology, 4250 Plymouth Rd., Ann Arbor, MI 48109, USA c Emory University School of Medicine, Atlanta, GA, USA d Atlanta VA Medical Center, Atlanta, GA, USA b

A R T I C L E I N F O

A B S T R A C T

Keywords: Sleep Prolonged exposure therapy Veterans Effectiveness Posttraumatic stress disorder

Sleep disturbances (SD) are pronounced in Veterans with posttraumatic stress disorder (PTSD). In clinical trials, SD have been shown to limit the effectiveness of evidence-based treatments for non-PTSD disorders. The purpose of this study was to investigate the relationships between pretreatment SD and the effectiveness of Prolonged Exposure (PE) therapy for Veterans with PTSD. Twenty-one Veterans completed the Pittsburgh Sleep Quality Index (PSQI) and the Clinician Administered PTSD Scale upon presenting to a PTSD specialty clinic. Veterans completed the PTSD Symptom Checklist-Civilian (PCL-C) at the initiation of PE and biweekly thereafter for the duration of treatment (96 total assessments). Correlations and hierarchical linear modeling were utilized to examine the potential impact of baseline sleep variables on the slope and magnitude of treatment outcomes. Higher PSQI total scores, and higher sleep latency and sleep medication use subscale scores were associated with higher PCL-C scores at baseline. Veterans evidenced significant reductions in PTSD symptoms during the course of the treatment study. Total PSQI scores and composites were not associated with reduced effectiveness of PE treatment or the slope of PTSD symptom changes. Sleep disturbances do not preclude Veterans from benefits derived from engagement in this gold standard PTSD intervention.

1. Introduction Sleep disturbances (SD), such as insomnia and nightmares, are hallmark symptoms of posttraumatic stress disorder (PTSD; American Psychiatric Association, 2013; Maher et al., 2006) and are particularly prominent among Veterans. For instance, among returning Veterans 30–67% report post-deployment SD (Hoge et al., 2007; Taylor et al., 2014). Among Veterans with PTSD, 44% endorse insomnia or nightmares in contrast with 6% of their peers without PTSD (Neylan et al., 1998). PTSD is further associated with increased risks for several sleepassociated medical disorders (i.e., obstructive sleep apnea, periodic limb movement disorder, restless legs syndrome, and pain; Maher et al.; Brownlow et al., 2015). Related, Bernardy et al. (2012) have described a prescription utilization rate of psychotropic medications for Veterans with PTSD approaching 80%. This is particularly salient, as each medication class examined has well-documented, and occasionally adverse, relationships with sleep (Maher et al., 2006). Researchers have posited that sleep onset and maintenance

⁎

problems in individuals with PTSD may be the result of increased hyperarousal and intrusive thoughts, concerns about loss of control while asleep, comorbid sleep disorders, fear of having nightmares, and/or decreased capacity for coping with PTSD-related distress due to lack of sleep (Harvey et al., 2003; Lamarche and De Konick, 2007). Moreover, Levin and Nielsen (2007) theorize that individuals with PTSD experience increased waking-state distress with heightened negative emotional reactivity, which may lead to an increased risk of nightmare formation. Research has shown such SD are predictive of the development of PTSD following a traumatic experience (Babson and Feldner, 2010). SD are further associated with greater PTSD severity, suicidal ideation, and are motivators of substance use (Bonn-Miller et al., 2014; Koren et al., 2002; Krakow et al., 2001; Nishith, 2001). Further, SD have been identified as potential threats to the effectiveness of evidence-based PTSD therapies by disrupting memory, limiting habituation, reducing daytime coping reserves, and impairing learning processes that are important in such interventions (Craske et al., 2008; Germain et al., 2008). Indeed, previous research on other

Corresponding author at: Ann Arbor Veterans Healthcare System, Mental Health Service (116 C), 2215 Fuller Rd., Ann Arbor, MI 48105, USA. E-mail address: [email protected] (M.B. Sexton).

http://dx.doi.org/10.1016/j.psychres.2017.06.044 Received 15 August 2016; Received in revised form 12 June 2017; Accepted 12 June 2017 Available online 14 June 2017 0165-1781/ Published by Elsevier Ireland Ltd.

Psychiatry Research 256 (2017) 118–123

M.B. Sexton et al.

et al., 1995, 2000) to confirm the presence of PTSD prior to referral to therapy and the Mini-International Neuropsychiatric Interview (MINI; Sheehan et al., 1998) to assess for comorbid psychiatric complaints. Upon enrollment in PE, Veterans completed the PTSD Checklist-Civilian version (PCL-C; Weathers et al., 1993) at their first session and every other week thereafter for the duration of the intervention. All Veterans who completed the PSQI at intake during this pilot study are included. Data was de-identified prior to analysis. Research was conducted in compliance with the Code of Ethics of the World Medical Association (Declaration of Helsinki). The Human Subjects Committee at the hospital approved the research protocol and procedures.

psychiatric disorders such as depression and social anxiety have found baseline sleep disturbances to be associated with inferior treatment outcomes (Buysse et al., 1999; Zalta et al., 2013). For example, baseline sleep quality in social anxiety disorder patients was inversely related to post-treatment social anxiety severity scores (Zalta et al., 2013). Likewise, in a sample of comorbid bipolar disorder and substance use disorder patients, those with greater baseline SD were less likely to experience reductions in mood symptoms immediately after treatment and at 6-month follow-up, although reductions in substance use were not associated with baseline SD (Putnins et al., 2012). Using objective assessments, Thase et al. (1996, 1997) observed individuals who did not remit following CBT for depression demonstrated abnormal EEG sleep profiles. Collectively, this literature suggests SD may impede recovery by reducing the efficacy of certain frontline psychotherapy treatments. If such a pattern extends to PTSD interventions, it may signal a need to address sleep complaints prior to the initiation of trauma-focused therapy. To-date, only one published study (Lommen et al., 2015) and one secondarily described non-published presentation (Kelly et al., 2015 as cited by Kelly et al., 2016) have investigated the potential influence of baseline SD on PTSD treatment outcome. Using cognitive therapy for PTSD (CT-PTSD) with a British clinical outpatient sample, Lommen et al. reported that pretreatment comorbid depressive symptoms were associated with a reduction in treatment gains. However, baseline SD did not limit the magnitude or speed of PTSD symptom reductions during the intervention and sleep and PTSD symptoms reduced during the intervention in tandem. These findings suggest the presence of SD do not preclude improvements in PTSD symptoms. Within the context of providing PTSD-related psychotherapy for U.S. Veterans, Clinical Practice Guidelines (2010) recommend the use of trauma-focused interventions that incorporate exposure and/or cognitive restructuring and the Veterans Healthcare System provides extensive specialty training in Prolonged Exposure (PE; Foa et al., 1991, 2007) therapy and Cognitive Processing Therapy (CPT; Resick and Schnicke, 1992) for use as frontline treatments for PTSD. Although the potential for SD to impede PE and CPT gains has not been investigated, the mechanisms proposed to explain the associations between SD and PTSD are directly targeted by exposure therapies such as PE (i.e. SD representing nighttime extensions of hyperarousal; fear of loss of control/inability to respond to danger; nightmares as conditioned stimuli; SD with arousals may limit habituation and memory processing; abnormal rapid eye movement (REM) impairing learning and emotional processing; SD-disturbances reducing daytime coping reserves; Harvey et al., 2003). Given the occurrence of concurrent SD and PTSD complaints within Veteran populations, it is imperative to evaluate whether the effectiveness of PE is limited by SD. Such information may improve our ability to ascertain if sleep-related interventions prior to or concurrent with PE are necessary in those with SD or whether it is preferable to engage Veterans in trauma-focused interventions as rapidly as possible to speed recovery and limit superfluous utilization of resources. To the best of our knowledge, there have not been any studies assessing baseline sleep characteristics as they relate to recovery following exposure-based treatments for PTSD. We aimed to test the hypotheses that 1) baseline SD would predict poorer treatment outcomes and 2) baseline SD would significantly slow the speed of treatment gains among Veterans presenting for trauma-focused psychotherapy.

2.1.1. Intervention Veterans enrolled in a course of PE (Foa et al., 2007), an empirically-supported and exposure-based intervention for PTSD. PE includes multiple techniques to address PTSD symptoms: imaginal exposure to the trauma memory and in vivo exposures to facilitate memory processing and reduce avoidance of feared or distressing stimuli, behavioral activation to promote engagement in valued activities, and breathing retraining. Staff and students trained by a VHA PE trainer conducted all treatments. Fidelity was not directly measured as part of standard care although providers attended weekly consultation and videotape/audiotape review was conducted with non-certified providers to promote protocol adherence. The recommended duration of PE varies by individual, response, and number of trauma memories addressed, but 8–14 weekly sessions are typically anticipated for a full course of treatment. 2.2. Measures 2.2.1. PTSD severity The PTSD Checklist-Civilian version (PCL-C; Weathers et al., 1993) is a 17-item self-report assessment of DSM-IV PTSD symptoms. All items are rated on a Likert scale from 1 (not bothered at all) to 5 (extremely bothered by symptom). Total scores range from 17 to 85 with scores above 50 considered a positive PTSD screen. 2.2.2. Sleep quality The Pittsburgh Sleep Quality Index (PSQI; Buysse et al., 1989) is a 19-item self-assessment survey of past-month sleep complaints. The PSQI is comprised of seven components: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, sleep medications, and daytime dysfunction. Each composite ranges from 0 to 3 with higher numbers suggestive of more problematic sleep quality in that domain. PSQI components are summed to create a PSQI Global Score (range = 0–21) and scores above five are considered reflective of a clinically significant sleep problem. The Global PSQI score has been shown to correlate well with other self-report non-physiological sleep assessments such as sleep logs (Buysse et al., 1989). The reliability and validity of the composite scores are less well established though may provide important contextual information. We present the primary results of the PSQI Global score and have elected to provide preliminary analyses on the domain scores with supplemental tables to assist in characterizing our sample and to facilitate future research. 2.3. Data analytic plan Participant characteristics are described with mean and percentage counts as warranted. Zero-order correlations were computed to assess the associations between variables. Despite the potential for overlap between the PSQI and the nightmare and insomnia items included on the PCL-C, we elected not to remove the sleep items from the PCL-C for several reasons. First, if sleep items change to a different degree or rate than other items on the PCL-C, such analyses may overestimate or underestimate the effectiveness of PE. Second, the psychometric qualities of the PCL-C with omitting specific items have not been

2. Materials and methods 2.1. Participants and procedures Participants included 21 Veterans engaged in Prolonged Exposure Therapy in a Midwestern Veteran Healthcare System PTSD Specialty Clinic in 2011. At intake, Veterans provided demographic information and completed the Clinician Administered PTSD Scale (CAPS; Blake 119

Psychiatry Research 256 (2017) 118–123

M.B. Sexton et al.

Table 1 Correlations among study variables.

(1) PCL-C Change Score (2) PCL-C Score Pre-Treatment (3) PCL-C Score Post-Treatment (4) PSQI Global Score (5) PSQI Sleep Quality (6) PSQI Sleep Latency (7) PSQI Sleep Duration (8) PSQI Sleep Efficiency (9) PSQI Sleep Disturbance (10) PSQI Use of Sleep Medication (11) PSQI Daytime Dysfunction

1

2

3

4

5

6

7

8

9

10

11

– – – – – – – – – – –

0.11 –

−0.82** 0.43 –

0.10 0.70** 0.31 – – – – – – – –

0.01 0.15 0.15 0.45* – – – – – – –

0.16 0.48* 0.10 0.64** −0.09 – – – – – –

0.19 0.60** 0.17 0.40 −0.11 0.28 – – – – –

0.13 0.15 −0.08 0.38 0.00 0.19 0.15 – – – –

0.03 0.21 0.05 0.42 0.11 0.25 −0.37 0.03 – – –

−0.23 0.39 0.44 0.48* 0.10 0.21 −0.19 −0.13 0.68** – –

−0.02 0.56* 0.38 0.75** 0.15 0.56** 0.42 −0.10 0.29 0.49* –

Notes. PCL-C = PTSD Checklist-Civilian, PSQI = Pittsburgh Sleep Quality Index. * p < 0.05. ** p < 0.01.

percent were male. The average participant was 47 years old (SD = 16). In terms of race, 85.7% of the sample self-identified as white, 4.8% as black, 4.8% as American Indian, and 4.8% declined to specify race. Regarding military and psychiatric characteristics, 86% of participants endorsed exposure to combat; 71% were seeking care related to combat exposure, 10% for military sexual trauma, and the remainder for nonmilitary trauma satisfying A1 criteria for the diagnosis of PTSD. Of Veterans for whom military era was available, 62% served during Vietnam and the remainder served during the Persian Gulf Era (including those enlisted in Operation Enduring Freedom/Operation Iraqi Freedom [OIF/OEF] efforts). Ten percent served during multiple era including wartime and peacetime eras. Characteristic of treatmentseeking Veterans with PTSD, psychiatric comorbidity was common and 76% had at least one other mental health disorder based on the MINI evaluation at intake. These included 62% with concurrent mood disorders, 33% with anxiety disorders, and 10% with alcohol and/or substance use disorders.

established. Finally, we wished to compare our research to other sleeprelated PTSD research that has incorporated the PSQI and unaltered PCL totals. Hierarchical linear modeling (HLM) was used to analyze the data and examine the potential impact of baseline sleep variables on treatment outcomes and the slope of the outcomes. HLM does not assume equal numbers of observations or measurements at fixed time points and missing data do not pose significant problems making it a practical analytic approach for treatment effectiveness data (Raudenbush and Bryk, 2002). More specifically, in the present study, HLM was used to examine longitudinal PCL-C outcomes (PCL-C total score; level-1) nested within patients (level-2) to assess response over time and overall treatment response. Variables that did not vary across time (i.e., sleep variables) were included at level-2. PCL-C total score was entered into the model as the outcome variable, week in treatment was entered at level-1, and sleep variables (i.e. Global PSQI score, PSQI component scores) were entered at level-2. A random intercept and time slope were included for each patient, which allowed estimation of the variance in PCL-C means (τβ00) as well as the variance in time effects (τβ11) across patients. Separate models were run in order to examine the impact of predictor variables on treatment response. There were a total number of ten models fitted in the present analyses. First, an unconditional model (i.e., Model 1) was fitted and intraclass correlations (ICCs) were calculated using sigma-squared (σ2) and tao (τ) variance estimates in order to determine the variance accounted for by each level of the data. Next, a model with week in treatment entered at level-1 (i.e., Model 2) was fitted to assess change in PTSD symptoms across time. Then, a model with week in treatment entered at level-1 and Global PSQI score entered at level-2 (i.e., Model 3) to assess whether baseline Global PSQI score impacted patient PTSD symptoms (PCL-C total score) during treatment was fitted. An interaction term was included to assess the impact sleep (as measured by PSQI scores) had on the slope of PCL-C scores across treatment. Similar separate models were then fitted for each of the PSQI component scores (i.e., week in treatment entered at level-1, PSQI component score entered at level-2, interaction term with week and sleep; Models 4–10). The variance of level-1 random errors and random effects at level-2 were estimated. All HLM analyses were conducted using the student version of HLM 7 software (Scientific Software Internation, Inc., Stokie, IL) and all other analyses were conducted using SPSS Statistics 22 (IBM Corporation, Armonk, NY). The a priori alpha for significance was established as p < 0.05 with two-tailed analyses.

3.2. Baseline trauma and sleep characteristics Prior to treatment, the average PCL-C score for Veterans was 61.7 (SD=11), suggestive of PTSD. The average PSQI was 11.1 (SD=3), exceeding the cut point suggestive of clinically significant sleep problems. 3.3. Zero-order correlations The associations among study variables are presented in Table 1. Pre-treatment PCL-C scores were significantly associated with global PSQI scores, the PSQI sleep latency composite, the PSQI sleep duration composite, and the PSQI daytime dysfunction composite. Change in PCL-C score across treatment (calculated by subtracting initial PCL-C score from the final PCL-C score for each patient) was not significantly correlated with any of the PSQI variables. HLM analyses of sleep impairments as predictors of PE treatment response and slope. Models of PTSD symptoms across treatment were based on 96 PCL-C measurements (level-1) nested within 21 patients (level-2). Model specifications and results are presented in Table 2 (Models 1–3) and Supplemental Table 1 (Models 4–10). Variance components for level-1 and level-2 were estimated via the unconditional model (i.e., Model 1). Results from this model indicated the presence of patient-level effects on treatment outcome (τβ00 = 122.85, χ2 = 114.78, p < 0.001). The ICC for between-patient variability was 0.53, indicating that 53% of the variance in PCL-C scores was accounted for by between person variability (i.e., factors associated with the patient). Model 2 (i.e., week in treatment entered at level-1) results indicated that week in treatment predicted PCL-C scores (t = −3.51, p = 0.002) such that PTSD

3. Results 3.1. Participant characteristics Twenty-one patients were included in the present study. Eighty-one 120

Psychiatry Research 256 (2017) 118–123

M.B. Sexton et al.

Table 2 Summary of HLM unconditional, week in treatment, and week and PSQI global score models (Models 1–3).

Fixed Effects: Level-1 Level-2 Random Effects: Level-1 Level-2 Deviance Parameters

Week Week*PSQI PSQI Intercept σ2 τ00 τ11

Model 1

Model 2

Model 3

Unconditional Model

Week in treatment added

Week & PSQI Global Score added

β

SE

t

β −1.31

SE 0.36

t −3.62**

55.38

2.67

20.75***

61.95

2.24

27.61***

Estimate 107.21 122.85

SD 10.35 11.08

χ2

Estimate 43.76 80.89 1.81 702.82 4

SD 6.62 8.99 1.34

χ2

749.62 2

114.78***

66.43*** 81.89***

β −0.79 −0.04 1.90 40.45

SE 1.76 0.14 0.65 7.53

t −0.45 −0.31 2.94** 5.37***

Estimate 42.52 52.13 2.07 695.84 4

SD 6.52 7.22 1.44

χ2 54.14*** 86.40***

Notes. HLM = Hierarchical Linear Model, PSQI = Pittsburgh Sleep Quality Index; β coefficient indicates amount of change in PCL-C scores per one unit change in variable (e.g., week in treatment, increase in PSQI global score); τ00 = estimation of the variance in PCL-C means across patients, τβ11 = estimation of the variance in time effects across patients; A deviance likelihood ratio test (chi-square) indicates whether including a random intercept or slope factor to each model significantly improves the fit of the overall model. * p < 0.05. ** p < 0.01. *** p < 0.001.

predict slope of PCL-C across treatment (see Supplemental Table 1). The PSQI Sleep Quality (Model 4), Sleep Duration (Model 6), Sleep Efficiency (Model 7), Sleep Disturbance (Model 8), and Daytime Dysfunction (Model 10) variables had no relationship with baseline PCL-C scores and did not significantly predict the slope of PCL-C scores across treatment or overall PCL-C scores (see Supplemental Table 1).

4. Discussion Consistent with previous literature (APA, 2013; Hoge et al., 2007; Maher et al., 2006; Taylor et al., 2014; Neylan et al., 1998), Veterans with elevated sleep complaints evidenced higher PTSD symptoms at treatment initiation. Further, HLM analyses revealed Veterans describing greater initiation insomnia and use of sleep medications reported more severe PTSD symptoms prior to treatment engagement. Specific to the effectiveness of PE, greater PTSD symptom severity at baseline was associated with higher PTSD severity posttreatment, though no sleep-related indicators were predictive of treatment effectiveness or the rate of improvements. While preliminary, these findings may offer some critical insight into appropriate treatment planning options for Veterans presenting for PTSD treatment with prominent SD and raise additional empirical questions. Presently, there is a lack of consensus regarding the ideal sequencing of interventions with this presentation (i.e. initiating trauma-focused treatment or cognitive behavioral therapy for insomnia [CBTI; Morin, 1993] first). Proponents of addressing sleep problems first often report that insomnia symptoms are present (though reduced) following PTSD therapy. However, the most often-cited study reporting this finding described participants engaging in several different types of cognitive-behavioral interventions for PTSD (Zayfert and DeViva, 2004). The sample size was small and it is unclear how many of these cases involved exposure-based interventions such as PE. In addition, we have noted that PTSD is often comorbid with other medical conditions that impair sleep (sleep apnea, pain, etc.). The use of single items from PTSD questionnaires likely does not differentiate between individuals with PTSD-associated nightmares and insomnia, which may be theoretically expected to decrease with evidence based treatments for PTSD, and medical conditions such as insomnia due to untreated obstructive sleep apnea or restless legs syndrome, which would not be anticipated to be responsive to either PTSD treatment or CBTI. More recently, an investigation of sleep changes produced within the context of CPT, PE or a minimal attention control group with 181 female sexual assault survivors found CPT and PE produced large

Fig. 1. PTSD Symptom Change by Session.

symptoms decreased by 1.31 points on the PCL-C per week in treatment. In addition, the week in treatment effect varied between patients, as indicated by significant τβ11 variance (τβ11 = 1.81, χ2 = 81.89, p < 0.001), suggesting that changes in PTSD symptoms across treatment varied for each patient. After accounting for overall treatment effects, additional models were fitted to examine the impact of baseline sleep variables on overall PCL-C scores as well as treatment outcomes (i.e., slope of PCL-C scores across treatment). With regard to the PSQI Global score, results indicated that greater overall sleep difficulties at baseline significantly predicted greater PTSD symptoms. More specifically, results suggested that for each one-point increase in the PSQI Global score, there was a 1.90-point increase in the PCL-C total score (t = 2.94, df =19, p = 0.008). No significant interaction was found (t = −0.31, p = 0.71), indicating that PSQI Global score did not predict slope of PCL-C across treatment. The average PCL-C scores over the course of care are illustrated in Fig. 1 (note: the figure is based on average scores by assessment session. As such, it may differ from HLM analyses that account for missingness in data). Separate models were fitted for each of the individual PSQI component scores (see Supplementary Table 1) and results indicated that only PSQI Sleep Latency (Model 5) and PSQI Use of Sleep Medication (Model 9) components had significant effects. More specifically, greater sleep latency predicted higher overall baseline PCL-C scores (t = 2.32, df =19, p = 0.03), but did not predict slope of PCL-C across treatment. Similarly, more frequent use of sleep medication predicted greater initial PCL-C scores (t = 2.99, df =19, p = 0.008), however, it did not 121

Psychiatry Research 256 (2017) 118–123

M.B. Sexton et al.

sleep complaints or are unwilling to engage in trauma-focused treatment. In these instances, protocols involving the incorporation of imagery rehearsal and CBTI appear the most promising (Casement et al.). While results of this study offer some positive insights, some limitations are noteworthy. The use of a clinical population lacks methodological rigor demonstrated in randomized and controlled designs. The results are restricted to changes in PTSD symptomatology and do not extend to investigating sleep-specific changes. The assessment utilized provides a gross estimation of sleep disturbances, but does not differentiate between PTSD-associated sleep complaints and those due to other etiologies. The use of specific medications that may alter sleep were not tracked and comorbidities were not controlled in this study. As such, it is possible that sleep medications may have reduced sleep symptoms concurrent with PE thus obfuscating the effect of sleep impairment on therapy outcomes. Future research incorporating these factors would be beneficial. Additionally, our preliminary findings are based on subjective report without direct observation of sleep or use of physiological assessments. This may be particularly salient as individuals complaining of insomnia or nightmares either with or without PTSD tend to self-report more substantial sleep impairments on the PSQI than are objectively observed. Related, elevated PSQI Global score ratings suggest greater perceived sleep disturbances but do not yield formal sleep diagnoses. Further research is warranted to ascertain if diagnosed sleep impairments alter PE effectiveness. While the PSQI Sleep Disturbances component evaluates several relevant disturbances suggestive of nightmares, OSA, nocturia, and pain, further research is needed to replicate our results and evaluate whether specific types of disturbances may be associated with PE outcomes. Finally, while the biweekly PCL-C administration approach utilized in this research increases the overall power of the design, it should be noted that the absolute sample size is small and relatively homogenous. As such, our findings should be considered preliminary and future research with larger sample sizes incorporating a control condition would permit more refined inclusion of potential sleep- or PTSD-related mediators that may influence results (the utilization of specific psychotropic medications, comorbid psychiatric and medical conditions that often co-occur with PTSD, etc.). Despite these weaknesses, this study is the first to evaluate the potential of SD to impede the effectiveness of PE, one of the primary psychotherapies offered by the Veterans Healthcare System to those experiencing PTSD. As SD are known to motivate Veterans to engage in care (Rosen et al., 2013), our results suggest that PE retains its’ value within the context of elevated self-reported sleep problems. The finding of the robustness of PE is particularly important given that evidencebased treatments for other disorders do appear diminished when SD are present as well as the presence of multiple comorbidities with demonstrated relationships to sleep. As such, our findings may reduce providers’ concerns about the utility of exposure-based trauma-focused care. This study highlights several additional gaps for future research. While SD does not appear to reduce the effectiveness of PE or CT-PTSD, research on other frontline treatments offered to Veterans (i.e. CPT, medications) would be illustrative. Further, comparative studies examining the results of trauma- and sleep-focused interventions on symptoms of PTSD, SD, functional impairment, and satisfaction with Veteran-specified goal attainments are warranted. Incorporations of physiological assessments and more thorough evaluation of the nature of sleep complaints (i.e. primary SD, PTSD-related SD, SD due to other medical conditions, SD satisfying diagnostic criteria for clinical sleep disorders) would be beneficial to extend our preliminary findings. In sum, our results replicate relationships between sleep disturbances and PTSD symptom severity. However, sleep disturbances did not impact the effectiveness of PE in reducing Veterans’ PTSD symptoms. Based on these findings, sleep disturbances should not be anticipated to preclude Veterans with PTSD from receiving and benefiting from this gold standard treatment.

reductions (d > 0.8) in nightmare and insomnia severity and moderate reductions (d > 0.5) on total PSQI scores and reports of sleep quality (Gutner et al., 2013). Demonstrated gains persisted through the longterm follow-up. The authors noted that a large percentage of participants continued to express clinically elevated sleep disturbances. However, details were not provided that might indicate if residual sleep symptoms may reflect those associated with medical complaints such as self-reported sleep complaints associated with pain, potential sleep apnea, or nocturia. Thus, it is difficult to ascertain whether persisting complaints reflected on PSQI scores were capturing sleep disturbances that would not be expected to be addressed by PTSD or insomnia treatments (i.e., untreated sleep apnea). CPT has also been associated with significant reductions in SD, though gains were not fully maintained through long-term follow up (Belleville et al., 2011). Other commonly articulated concerns are that trauma-focused treatments are emotionally taxing, anxiety provoking, and prone to result in premature attrition (Baddeley and Gros, 2013; Schottenbauer et al., 2008). Thus, it has been argued CBTI utilized as a ‘preparatory intervention’ for Veterans presenting for PTSD-related care may offer a stabilization advantage that can motivate and ease patients into trauma-focused treatments after successful completion of CBTI. Yet, a comprehensive review specific to adherence to CBTI suggests 14–40% of participants discontinue before mid-treatment assessments: a rate similar to drop from trauma-focused therapy (Matthews et al., 2013). If such patterns extend to Veterans with PTSD, we can anticipate that missed opportunities to engage individuals in frontline PTSD treatments as soon as possible may result, not only in a delay of recovery, in an increased likelihood that trauma-focused interventions are never initiated. Rather, initial focus on the treating the primary complaint (i.e. PTSD that presents with SD) may more adequately address the full spectrum of symptom complaints. Our preliminary results suggest PE retains its effectiveness in the context of sleep disturbances, though replication with gold standard sleep measurements is warranted. In addition, future research assessing whether concurrent and integrated PE with CBTI adds to the efficacy of results beyond PE alone may be beneficial for assisting with treatment planning. For instance, recent research utilizing hypnosis prior to CPT augmented sleep and depression, though not PTSD, symptom reduction (Galovski et al., 2016). Future studies that assess sleep and PTSD symptoms during the course of PE would be beneficial to evaluate whether sleep changes are temporally associated with PTSD symptom reduction (i.e. if sleep improvements may precede, follow, or occur contemporaneous to PTSDrelated changes). Our finding that sleep disturbances did not reduce the effectiveness of PE is consistent with Lommen et al.’s recent study (2015), which found SD did not directly reduce the usefulness of cognitive therapy for PTSD, except in instances with elevated depression. Together, these findings suggest the utility of trauma-focused treatments is not reduced when sleep disturbances are present at baseline. As such, it does not appear necessary to engage in a sleep treatment as a preparatory or stabilizing step prior to engagement evidence-based treatment for PTSD in order to obtain benefits, though additional research is required to replicate our preliminary results. It may be the case that motivating and supporting Veterans to engage in trauma-focused treatments and reserving CBTI for more treatment refractory SD may provide the greatest advantages for clients, particularly in cases where engagement in multiple, sequential interventions is not desired or is unlikely to occur. As noted, further inquiry examining the potential for added benefits of integrated PE and CBTI may further assist with tailoring care recommendations for maximal outcomes. In addition, previous research has found sleep-specific treatments do reduce core symptoms of PTSD (Swanson et al., 2009; Margolies et al., 2013; Ulmer et al., 2011; see also reviews by Casement and Swanson, 2012; Nappi et al., 2012), albeit the effect sizes may be less robust than for PE (Powers et al., 2010). Continued research to refine, dismantle, and amplify sleep-focused interventions is warranted for Veterans that may be primarily focused on 122

Psychiatry Research 256 (2017) 118–123

M.B. Sexton et al.

Hoge, C.W., Terhakopian, A., Castro, C.A., Messer, S.C., Engel, C.C., 2007. Association of posttraumatic stress disorder with somatic symptoms, health care visits, and absenteeism among Iraq war veterans. Am. J. Psychiatry 164, 150–153. Kelly, M.R., Killgore, W.D.S., Haynes, P.L., 2016. Understanding recent insights in sleep and posttraumatic stress disorder from a research domain criteria (RDoC) framework. Curr. Sleep Med. Rep. 2 (4), 223–232. Koren, D., Arnon, I., Lavie, P., Klein, E., 2002. Sleep complaints as early predictors of posttraumatic stress disorder: a 1-year prospective study of injured survivors of motor vehicle accidents. Am. J. Psychiatry 159 (5), 855–857. Krakow, B., Germain, A., Warner, T.D., Schrader, R., Koss, M., Hollifield, M., Johnston, L., 2001. The relationship of sleep quality and posttraumatic stress to potential sleep disorders in sexual assault survivors with nightmares, insomnia, and PTSD. J. Trauma Stress 14 (4), 647–665. Lamarche, L.J., De Koninck, J., 2007. Sleep disturbance in adults with posttraumatic stress disorder: a review [CME]. J. Clin. Psychiatry 68 (8), 1257–1270. Levin, R., Nielsen, T.A., 2007. Disturbed dreaming, posttraumatic stress disorder, and affect distress: a review and neurocognitive model. Psychol. Bull. 133 (3), 482. Lommen, M.J., Grey, N., Clark, D.M., Wild, J., Stott, R., Ehlers, A., 2015. Sleep and treatment outcome in posttraumatic stress disorder: results from an effectiveness study. Depress. Anxiety. Maher, M.J., Rego, S.A., Asnis, G.M., 2006. Sleep disturbances in patients with posttraumatic stress disorder: epidemiology, impact and approaches to management. CNS Drugs 20 (7), 567–591. Margolies, S.O., Rybarczyk, B., Vrana, S.R., Leszczyszyn, D.J., Lynch, J., 2013. Efficacy of a cognitive‐behavioral treatment for insomnia and nightmares in Afghanistan and Iraq veterans with PTSD. J. Clin. Psychol. 69 (10), 1026–1042. Matthews, E.E., Arnedt, J.T., McCarthy, M.S., Cuddihy, L.J., Aloia, M.S., 2013. Adherence to cognitive behavioral therapy for insomnia: a systematic review. Sleep Med. Rev. 17 (6), 453–464. Morin, C.M., 1993. Insomnia: Psychological Assessment and Management. The Guilford Press, New York. Nappi, C.M., Drummond, S.P., Hall, J.M., 2012. Treating nightmares and insomnia in posttraumatic stress disorder: a review of current evidence. Neuropharmacology 62 (2), 576–585. Neylan, T.C., Marmar, C.R., Metzler, T.J., Weiss, D.S., Zatzick, D.F., Delucchi, K.L., et al., 1998. Sleep disturbances in the Vietnam generation: findings from a nationally representative sample of male Vietnam veterans. Am. J. Psychiatry 155 (7), 929–933. Nishith, P., Resick, P.A., Mueser, K.T., 2001. Sleep difficulties and alcohol use motives in female rape victims with posttraumatic stress disorder. J. Trauma. Stress 14 (3), 469–479. Powers, M.B., Halpern, J.M., Ferenschak, M.P., Gillihan, S.J., Foa, E.B., 2010. A metaanalytic review of prolonged exposure for posttraumatic stress disorder. Clin. Psychol. Rev. 30 (6), 635–641. Putnins, S.I., Griffin, M.L., Fitzmaurice, G.M., Dodd, D.R., Weiss, R.D., 2012. Poor sleep at baseline predicts worse mood outcomes in patients with co-occurring bipolar disorder and substance dependence. J. Clin. Psychiatry 73 (5), 703–708. Raudenbush, S.W., Bryk, A.S., 2002. Hierarchical Linear Models: Applications and Data Analysis Methods Vol. 1 Sage, Thousand Oaks, CA. Resick, P.A., Schnicke, M.K., 1992. Cognitive processing therapy for sexual assault victims. J. Consult. Clin. Psychol. 60 (5), 748–756. Rosen, C., Tiet, Q., 2013. Presenting concerns of veterans entering treatment for posttraumatic stress disorder. J. Trauma. Stress 26 (5), 640–643. Schottenbauer, M.A., Glass, C.R., Arnkoff, D.B., Tendick, V., Gray, S.H., 2008. Nonresponse and dropout rates in outcome studies on PTSD: review and methodological considerations. Psychiatry 71 (2), 134–168. Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorium, P., Janavs, J., Weiller, E., et al., 1998. The Mini-International Neuropsychiatric Interview (M.I.N.I): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD10. J. Clin. Psychiatry 59 (suppl 20), 22–33. Swanson, L.M., Favorite, T.K., Horin, E., Arnedt, J.T., 2009. A combined group treatment for nightmares and insomnia in combat veterans: a pilot study. J. Trauma. Stress 22 (6), 639–642. Taylor, M.K., Hilton, S.M., Campbell, J.S., Beckerley, S.E., Shobe, K.K., Drummond, S.P., 2014. Prevalence and mental health correlates of sleep disruption among military members serving in a combat zone. Mil. Med. 179 (7), 744–751. Thase, M.E., Simons, A.D., Reynolds, C.F., 1996. Abnormal electroencephalographic sleep profiles in major depression: association with response to cognitive behavior therapy. Arch. Gen. Psychiatry 53 (2), 99–108. Thase, M.E., Buysse, D.J., Frank, E., Cherry, C.R., Cornes, C.L., Mallinger, A.G., Kupfer, D.J., 1997. Which depressed patients will respond to interpersonal psychotherapy? The role of abnormal EEG sleep profiles. Am. J. Psychiatry 154 (4), 502–509. The Management of Post-Traumatic Stress Working Group, 2010. VA/DoD Clinical Practice Guideline for Management of Post-traumatic Stress. 〈http://www. healthquality.va.gov/guidelines/MH/ptsd/cpgPTSDFULL201011612c.pdf〉. Ulmer, C.S., Edinger, J.D., Calhoun, P.S., 2011. A multi-component cognitive-behavioral intervention for sleep disturbance in veterans with PTSD: a pilot study. J. Clin. Sleep Med. 7 (1), 57–68. Weathers, F.W., Litz, B.T., Herman, D.S., Huska, J.A., Keane, T.M., 1993. The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility. In: Annual Convention of the International Society for Traumatic Stress Studies. International Society for Traumatic Stress Studies, San Antonio. Zalta, A.K., Dowd, S., Rosenfield, D., Smits, J.A., Otto, M.W., Simon, N.M., Pollack, M.H., 2013. Sleep quality predicts treatment outcome in CBT for social anxiety disorder. Depress. Anxiety 30 (11), 1114–1120. Zayfert, C., DeViva, J.C., 2004. Residual insomnia following cognitive behavioral therapy for PTSD. J. Trauma. Stress. 17 (1), 69–73.

Disclosure/Conflict of interest Dr. Arnedt holds a NIH grant (R42 MD008845-02) through Zansors, LLC. The authors declare no additional financial disclosures or conflicts of interest. Role of funding Research was supported by the Mental Health Service at Ann Arbor Veterans Healthcare System. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgements The authors thank Margaret Venners, MPH, MSW1 and Mahrie E. Defever, BA for database management. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.psychres.2017.06.044. References American Psychiatric Association, 2013. Diagnostic and Statistical Manual Of Mental Disorders (DSM-5®). American Psychiatric Pub., Arlington, VA. Babson, K.A., Feldner, M.T., 2010. Temporal relations between sleep problems and both traumatic event exposure and PTSD: a critical review of the empirical literature. J. Anxiety Disord. 24 (1), 1–15. Baddeley, J.L., Gros, D.F., 2013. Cognitive behavioral therapy for insomnia as a preparatory treatment for exposure therapy for posttraumatic stress disorder. Am. J. Psychother. 67 (2), 199–210. Belleville, G., Guay, S., Marchand, A., 2011. Persistence of sleep disturbances following cognitive-behavior therapy for posttraumatic stress disorder. J. Psychosom. Res. 70 (4), 318–327. Bernardy, N.C., Lund, B.C., Alexander, B., Friedman, M.J., 2012. Prescribing trends in veterans with posttraumatic stress disorder. J. Clin. Psychiatry 73 (3), 297–303. Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G., Gusman, F.D., Charney, D.S., Keane, T.M., 1995. The development of a clinician‐administered PTSD scale. J. Trauma. Stress 8 (1), 75–90. Blake, D.D., Weathers, F.W., Nagy, L.M., Kaloupek, D.G., Charney, D.S., Keane, T.M., 2000. The Clinician Administered PTSD scale (CAPS) for DSM-IV (Computer Software and Manual). Dept. Vet. Affairs, Boston, MA. Bonn-Miller, M.O., Babson, K.A., Vandrey, R., 2014. Using cannabis to help you sleep: heightened frequency of medical cannabis use among those with PTSD. Drug Alcohol Depen. 136, 162–165. Brownlow, J.A., Harb, G.C., Ross, R.J., 2015. Treatment of sleep disturbances in posttraumatic stress disorder: a review of the literature. Curr. Psychiatry Rep. 17 (6), 1–10. Buysse, D.J., Reynolds, C.F., Monk, T.H., Berman, S.R., Kupfer, D.J., 1989. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res. 28 (2), 193–213. Buysse, D.J., Tu, X.M., Cherry, C.R., Begley, A.E., Kowalski, J., Kupfer, D.J., Frank, E., 1999. Pretreatment REM sleep and subjective sleep quality distinguish depressed psychotherapy remitters and nonremitters. Biol. Psychiatry 45 (2), 205–213. Casement, M.D., Swanson, L.M., 2012. A meta-analysis of imagery rehearsal for posttrauma nightmares: effects on nightmare frequency, sleep quality, and posttraumatic stress. Clin. Psychol. Rev. 32 (6), 566–574. Craske, M.G., Kircanski, K., Zelikowsky, M., Mystkowski, J., Chowdhury, N., Baker, A., 2008. Optimizing inhibitory learning during exposure therapy. Behav. Res. Ther. 46 (1), 5–27. Foa, E.B., Rothbaum, B.O., Riggs, D.S., Murdock, T.B., 1991. Treatment of posttraumatic stress disorder in rape victims: a comparison between cognitive-behavioral procedures and counseling. J. Consult. Clin. Psychol. 59 (5), 715. Foa, E.B., Hembree, E.A., Rothbaum, B.O., 2007. Prolonged Exposure Therapy for PTSD. Oxford University, New York. Galovski, T.E., Harik, J.M., Blain, L.M., Elwood, L., Gloth, C., Fletcher, T.D., 2016. Augmenting cognitive processing therapy to improve sleep impairment in PTSD: a randomized controlled trial. J. Consult. Clin. Psychol. 84 (2), 167. Germain, A., Buysse, D.J., Nofzinger, E., 2008. Sleep-specific mechanisms underlying posttraumatic stress disorder: integrative review and neurobiological hypotheses. Sleep Med. Rev. 12 (3), 185–195. Gutner, C.A., Casement, M.D., Gilbert, K.S., Resick, P.A., 2013. Change in sleep symptoms across cognitive processing therapy and prolonged exposure: a longitudinal perspective. Behav. Res. Ther. 51 (12), 817–822. Harvey, A.G., Jones, C., Schmidt, D.A., 2003. Sleep and posttraumatic stress disorder: a review. Clin. Psych. Rev. 23 (3), 377–407.

123