Posttraumatic stress disorder (PTSD) is a condition that can emerge after exposure to a traumatic event. PTSD is characterized by several symptoms, including intrusion symptoms associated with the traumatic event: distressing memories or dreams, and/or dissociative reactions; psychological distress at exposure to trauma cues; physiologic reactions to trauma cues; avoidance of stimuli associated with the traumatic event; negative alterations in cognitions and mood associated with the trauma; and alterations in arousal and reactivity, including sleep disturbance. The purpose of the systematic review was to synthesize the evidence from randomized controlled trials (RCTs) on the effects of interventions for adults with PTSD on sleep outcomes.
The review was guided by the following key questions (KQs) and subquestions:
- KQ1. What are the effects of interventions in PTSD on sleep outcomes (primary outcome), PTSD symptoms, and adverse events (AEs)?
- KQ1a. Do the effects differ by type of intervention?
- KQ1b. Do the effects differ by trauma type?
- KQ1c. Do the effects differ by treatment setting and modality?
We searched research databases—the Allied and Complementary Medicine Database, the Cochrane Central Register of Controlled Trials, the Cumulative Index to Nursing and Allied Health Literature, Embase, Published International Literature on Traumatic Stress, PubMed, and PsycInfo—in October 2019, as well as bibliographies of existing systematic reviews and included studies, to identify pertinent trials published in English. Two reviewers independently screened literature identified by the searches using predetermined eligibility criteria, abstracted information, and outcome data from those studies that met the inclusion criteria, and they assessed the risk of bias. The primary outcome domain included sleep quality and sleep disturbances such as insomnia and nightmares. Secondary outcomes were PTSD symptoms and AEs of the interventions. Interventions of interest included pharmacological, psychological, and behavioral treatments, or complementary and integrative medicine treatments, aimed at improving sleep or lessening other PTSD symptoms.
Data abstraction was undertaken by one reviewer and checked by another experienced reviewer with an online database designed for systematic reviews. The risk of bias was assessed using the Cochrane Risk of Bias tool. Study results were converted to effect sizes for the outcomes of interest. Results across studies were summarized in Hartung-Knapp corrected random effects meta-analysis where data allowed, and all results were described in a narrative synthesis. We conducted preplanned subgroup analyses and meta-regressions for intervention subgroups, differentiating interventions that explicitly aim to improve sleep from treatments for PTSD in general that report an effect on sleep. In addition, we differentiated by trauma type (origin and context of trauma) and explored differences in treatment setting and modality.
Sensitivity analyses, such as excluding high-risk-of-bias studies or using trim-and-fill methods where evidence of publication bias was detected, were used to assess the robustness of study results. We assessed the quality of evidence for each outcome using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. The review protocol was registered in PROSPERO (identification number: CRD42018102200).
The identified interventions included pharmacological, psychological, and behavioral treatments, or complementary and integrative medicine treatments, aimed at improving sleep or lessening other PTSD symptoms in patients diagnosed with PTSD.
KQ1. Effects of Interventions
Across studies, the interventions showed a medium effect on sleep (standardized mean difference [SMD] −0.56; 95-percent confidence interval [CI] −0.75, −0.37; 49 RCTs). Regarding specific sleep outcomes of interest, the interventions showed a medium effect on the Pittsburgh Sleep Quality Index, or PSQI (SMD −0.51; CI −0.76, −0.25; 29 RCTs), and interventions also had a medium effect on the Clinician-Administered PTSD Scale (CAPS) sleep quality item (SMD −0.51; CI −0.96, −0.06; seven RCTs).
Studies used a variety of insomnia measures, and we also found improvements regarding insomnia (SMD −0.65; CI −1.01, −0.28; 17 RCTs) across all measures and also when restricting to studies reporting on the Insomnia Severity Index, or ISI (SMD −0.73; CI −1.36, 0.11; ten RCTs). In addition, studies reporting on nightmare measures also indicated improvements (SMD −0.56; −0.99, −0.12; 15 RCTs), including studies that used the CAPS nightmare item (SMD −0.55; CI −0.94; −0.17; 12 RCTs).
The interventions also showed a medium effect on PTSD symptoms across outcomes (SMD −0.48; CI −0.67, −0.29; 44 RCTs). There was a medium effect on CAPS (SMD −0.38; CI −0.64, −0.13; 17 RCTs) and the PTSD Checklist, or PCL (SMD −0.50; CI −0.81, −0.19; 11 RCTs).
Most AEs did not differ significantly between interventions and passive comparators, but we found more incidences of diarrhea and dizziness associated with predominantly pharmacological associations.
KQ1a. Do the Effects Vary by Type of Intervention?
Studies evaluated a range of medication, psychotherapy, and complementary and alternative medicine interventions aimed at improving sleep and lessening PTSD symptoms. We identified a number of studies comparing two alternative interventions. However, no two studies compared the same interventions, and all evaluations were unique and not replicated in another study.
We detected considerable heterogeneity, and that the effects of individual studies varied. We explored differences between interventions in indirect comparisons across studies. Interventions explicitly aimed at improving sleep tended to report larger effects on sleep than interventions aimed at treating PTSD without an explicit sleep focus (SMD −0.78; CI −1.12, −0.44; I2 81 percent; 24 RCTs; versus SMD −0.36; CI −0.54, −0.17; I2 62 percent; 25 RCTs; p = 0.03). Furthermore, sleep-focused psychotherapy interventions appeared to report larger effects on sleep than did other interventions (SMD −1.40; CI −2.40, −0.40; four RCTs; p = 0.02). The differences between interventions should be interpreted with caution because they are based on indirect comparisons across studies and not direct evidence from head-to-head comparisons.
KQ1b. Do the Effects Differ by Trauma Type?
The included studies assessed PTSD in a variety of samples, including 40 studies in military or veteran samples. We also explored the effect of trauma type on sleep, but we did not identify systematic differences between military and nonmilitary trauma populations (p = 0.35).
KQ1c. Do the Effects Differ by Treatment Setting and Modality?
One-third of the studies offered interventions in specialty care, while others included other settings; we did not detect systematic differences on sleep outcomes in indirect comparisons across studies (p = 0.76). Differentiating outpatient and inpatient settings also did not identify systematic differences in sleep effects. Finally, the large majority of studies evaluated individual treatments, and we detected no systematic differences between individual and group approaches in meta-regressions (p = 0.95). We found only one study comparing treatment modalities, but the lack of difference between modalities was not replicated in another independent study.
Analyzing studies that report on sleep outcomes in patients with PTSD, we found that interventions aimed at improving sleep or lessening other PTSD symptoms do, in fact, improve sleep. Interventions specifically targeting sleep—in particular, psychotherapy specifically addressing sleep—tended to report larger effects on sleep outcomes. Treatment effect estimates varied across studies, and heterogeneity in results, was substantial. However, effects on sleep were not systematically affected by whether trauma was military or nonmilitary in origin, and we detected no systematic differences of treatment settings and modalities. Comparative effectiveness studies are needed to support the findings.