Psychotherapy Trials

This appendix provides information on trial findings about psychotherapies (Key Question [KQ] 1) in the same basic order as the psychological interventions covered in the Results chapter – namely, cognitive behavioral interventions of all types, eye movement desensitization and reprocessing (EMDR), and other psychotherapies. Several of these various trials did have findings consistent with those reported in the main systematic review.

Tables in this appendix describe the trials. Entries are in alphabetical order by last name of the first author of the article(s) in question. For duration of treatment (or trial), in some cases the investigators reported only the number of sessions, rather than the length of the treatment or the trial (in, e.g., weeks or months). Some trials measured certain outcomes at some point, such as 3 months, after treatment ended (usually referred to as posttreatment). Disease severity was measured frequently by the Clinician-Administered PTSD Scale (CAPS); these are the data reported in the tables unless another instrument, such as the Davidson Trauma Scale or the PTSD Checklist, is specified. Severity measures can be reported as mean scores or a range of mean scores across groups for the CAPS or another PTSD measure listed. As in the main report, inactive arms for these trials included waitlist or usual care (which included treatment as usual as well).

Pharmacologic Trials

We also present information on the high risk-of-bias trials of pharmacological interventions (KQ 2). These include a wide array of antidepressants, anticonvulsants (or mood stabilizers), and antipsychotics, among other pharmacologic agents. As with psychological interventions, we report on both efficacy trials (placebo controls) and on comparative effectiveness (from head-to-head trials – i.e., active comparators). Several of these various trials did have findings consistent with those reported in the main systematic review.

Tables in this appendix describe the trials using the organization previously described.

Cognitive Behavioral Therapy

Several trials addressed a wide array of cognitive behavioral therapies (CBT). These included cognitive interventions, coping skills interventions, exposure treatments of various types, and interventions with “mixed” CBT components, including various combinations of psychoeducation, self-monitoring, stress management, relaxation training, skills training, exposure (imaginal, or in vivo, or both), cognitive restructuring, guided imagery, mindfulness training, breathing retraining, crisis/safety planning, and relapse prevention.

Three separate trials tested three cognitive interventions: two tested cognitive processing therapy (CPT) and one tested web-based cognitive therapy (CT) (see Table G-1). For the CPT trial with person-centered therapy (PCT) as the comparator, we considered PCT to be an active comparator but not one for which we were interested in examining the comparative effectiveness with interventions of interest. For the other CPT trail and the CT trial, both provided findings consistent with those from the trials described in the main report on all outcomes.

We rated three trials of CBT- coping skills as high risk of bias (Table G-2). One trial compared stress inoculation training (SIT) with narrative exposure therapy (NET); another compared SIT with a waitlist group. The third trial tested an affect management intervention versus waitlist. Treatment-related improvements in reducing PTSD symptoms at posttreatment were reported for SIT versus waitlist only. These findings are consistent with the single trials of these interventions included in our qualitative synthesis.

We rated 17 trials of CBT - exposure interventions (see Table G-3). The types of “exposure” therapies differed appreciably across these trials. Like the included studies, a majority of the high risk-of-bias trials of exposure found that the CBT-exposure group did significantly better than the inactive comparator group, especially with respect to reducing PTSD symptoms (using any of the available symptom measures), achieving remission, and losing the PTSD diagnosis. Persons in the exposure interventions also tended to do better as well in reducing depression symptoms. Although these findings were similar to those from the low or medium risk-of-bias trials, precision was lower, with wider confidence intervals. Similar to the set of included trials, few high risk-of-bias trials examined anxiety symptoms, functional status, quality of life, and functional impairment outcomes; when they did, however, findings were similar to those of included studies.

Fifteen trials tested a considerable array of “mixed” CBT interventions (Table G-4). Generally, their findings were consistent with those from low or medium risk-of-bias trials on all the outcomes examined. Specifically, a majority of the high risk-of-bias studies found that the CBT- mixed group had significantly better results than did those in the various comparison groups; these included reducing PTSD symptoms, achieving remission, and losing the PTSD diagnosis, as well as reducing depression symptoms. Similar to the set of included study evidence, few of these omitted trials examined anxiety symptoms, substance use, functional status, quality of life or functional impairment outcomes; when they reported such findings, however, they were similar to those from included studies.

Eye Movement Desensitization and Reprocessing

Seven trials of eye movement desensitization and reprocessing (EMDR) were rated high risk of bias (see Table G-5). These seven trials had findings consistent with those from included trials on various outcomes. Like the trials rated either low or medium risk of bias, a majority of the high risk-of-bias studies found that the EMDR group had significantly better outcomes than patients in an inactive comparator group in terms of reducing PTSD symptoms, losing the PTSD diagnosis, and reducing depression. Few of these seven trials examined any other coexisting condition, functional status, quality of life, or disability or functional impairments.

Analyses of Other Psychological Interventions

Ten trials that we rated high risk of bias dealt with a variety of other psychological interventions (Table G-6). Three of these assessed narrative exposure therapy (NET); of these three, the one comparing NET with usual care found results consistent with those from medium risk-of-bias trials. The other two NET trials had active comparators -- stress inoculation therapy and psychoeducation. Whereas the IPT study rated high risk of bias tested the efficacy of IPT versus treatment as usual, the IPT study included in the main study findings compared the effectiveness of IPT versus two other treatments, precluding comparisons of the findings. Other studies rated high risk of bias assessed interventions (or comparators) not included for the main analyses, namely slow breathing (SB) feedback, and trauma-focused CBT (TF-CBT) plus breathing biofeedback; thus, we could not assess, for example, the consistency of the results.

Alpha Blockers

We rated two trials as high risk of bias (Table G-8). Both compared prazosin with placebo.

Anticonvulsants

We rated three placebo-controlled trials as high risk of bias (Table G-9). the interventions included one trial each of divalproex, lamotrigine, and topiramate.

Atypical Antipsychotics

We found no trials of atypical antipsychotics versus placebo that met our eligibility criteria but could be rated as either low or medium risk of bias. Four trials, however, were eligible but rated high risk of bias (Table G-10). Two trials tested risperidone, and one each tested aripiprazole or quetiapine,

Benzodiazepines

A single trial in this drug class tested alprazolam against placebo (Table G-11). As noted in the main report, no benzodiazepine trial was rated as either low or medium risk of bias. Evidence is insufficient to determine the efficacy of benzodiazepines for improving outcomes for adults with PTSD.

Selective Serotonin Reuptake Inhibitors

Among the trials for SSRIs, one studied fluoxetine and one examined paroxetine (both against placebo) (Table G-12). A third was more complicated: sertraline combined with mirtazapine (a newer type of antidepressant in the class known as tetracyclic piperazinoazepine) against sertraline combined with placebo. Findings reported for these trials were generally consistent with what investigators found in low or medium risk-of bias trials.

Other Second-Generation Antidepressants

We rated one trial comparing nefazodone (a phenylpiperazine antidepressant) with placebo as high risk of bias (Table G-13).

One trial of nefazodone and an active comparator – sertraline (an SSRI) – was rated high risk of bias (Table G-14).

Tricyclic Antidepressants

We rated three trials of otherwise meeting criteria for this section as high risk of bias (Table G-15). All were placebo-controlled trials conducted more than 20 years ago: one of amitriptyline, one of imipramine, and one of desipramine.