Key Points: CBT—Cognitive Interventions

• Moderate SOE supports the efficacy of CPT and CT for reduction in PTSD symptoms, loss of PTSD diagnosis, and reduction in depression symptoms
• Moderate SOE supports the efficacy of CT for reduction in anxiety symptoms and reduction in disability.

Key Points: CBT—Coping Skills

• Insufficient evidence exists to determine the efficacy of relaxation, stress inoculation training (SIT), and structured approach therapy (SAT), with single studies testing each intervention.
• Moderate SOE supports the effectiveness of CBT-exposure compared with relaxation for reduction in PTSD symptoms, loss of PTSD diagnosis, and reduction in depression symptoms
• Low SOE supports the effectiveness of CBT-mixed compared with relaxation for reduction in PTSD symptoms.

Key Points: CBT—Exposure

• High SOE supports the efficacy of CBT-exposur etherapy for reduction in PTSD symptoms, loss of PTSD diagnosis, and reduction in depression symptoms, and low SOE for anxiety symptoms.
• Moderate SOE provides comparative effectiveness of CBT-exposure compared with relaxation for reduction in PTSD symptoms and loss of PTSD diagnosis, and low SOE favors CBT-exposure over relaxation for reduction in depression symptoms.
• Low SOE shows no difference in effectiveness between CBT-exposure and EMDR for reduction in PTSD symptoms and between CBT-exposure and CBT-exposure+CR for reduction in depression symptoms.

Key Points: CBT—Mixed

• High SOE supports the efficacy of CBT-mixed for reduction in PTSD symptoms, loss of PTSD diagnosis, and reduction in depression symptoms.
• Moderate SOE supports the efficacy of CBT-mixed for reduction in anxiety symptoms and reduction in substance use issues.
• Low SOE supports the efficacy of CBT-mixed for reduction in disability/functional impairment.
• Low SOE supports the comparative effectiveness of CBT-mixed over relaxation for reduction in PTSD symptoms.

Key Points: EMDR

• Moderate SOE supports the efficacy of EMDR for reduction in PTSD symptoms, loss of diagnosis, and reduction in depressive symptoms.
• Low SOE shows no difference in effectiveness for reduction in PTSD symptoms between EMDR and CBT-exposure.

Key Points: Other Psychological Interventions

• Low SOE supports the efficacy of TAR for reduction in PTSD symptoms.
• Low SOE supports the efficacy of BEP for loss of PTSD diagnosis, reduction in depression symptoms, and reduction in anxiety symptoms.
• Low SOE supports efficacy of IRT for reduction in PTSD symptoms.
• Moderate SOE supports the efficacy of NET for reduction in PTSD symptoms and low SOE supports its efficacy for loss of PTSD diagnosis. Low SOE supports no difference of SS versus inactive comparator on reduction in PTSD symptoms.

Characteristics of Studies

Table 7 summarizes the characteristics of the 14 cognitive intervention studies that met our inclusion criteria. We divided the 14 cognitive interventions further into CPT (7 studies), CR (1 study), CT (5 studies), and meta cognitive therapy (MCT) (1 study).

Table 7

Characteristics of included cognitive intervention trials.

Of the seven CPT interventions, four had wait-list comparators,^{1–3, 6} one had a usual-care comparator,⁴ and three had active intervention comparators (prolonged exposure [PE],³ MEST,¹²⁴ and group PCT).¹²⁷ The single CR study had three active comparators: relaxation, PE, and a combined CR and PE.¹²² Five studies tested CT: three had wait-list comparators,^{5, 8, 9} one had a usual care comparator,⁷ and two had active comparators⁹ (self-help booklet⁵ and imaginal exposure¹²⁹). The single MCT study had one active comparator, PE, and one inactive comparator, wait-list.¹⁹ Further details describing the included studies are provided in Appendix F.

Of the seven CPT studies, sample sizes ranged from 16 to 171. Duration of treatment ranged from 6 to 17 weeks. Five studies included at least one posttreatment followup assessment after 1 to 12 months,^{1–4, 124} one study reported followup data but had a cross-over design affecting time period comparisons across groups,⁶ and one study reported outcomes at 5 to 10 years followup.¹²⁶ Three studies enrolled all or a majority of females with sexual abuse or assault trauma types,^{2, 3, 6} and three studies enrolled all or a majority of males with combat-related trauma types.^{1, 4, 127} The mean age of participants in the CPT studies ranged from 32 to 54 years. The primary outcome measures for these studies were the Clinician-Administered PTSD Scale (CAPS), PTSD Checklist (PCL), Modified PTSD Symptom Scale (MPSS), and PTSD Symptom Scale (PSS).

The single CR study that contained three active comparators included males and females with exposure to mixed trauma types.¹²² Participants completed 10 sessions over a mean of 16 weeks and had followup assessments at 1, 3, and 5 months posttreatment. The primary outcome measure was the Impact of Event Scale (IES).

Of the five CT studies, sample sizes ranged from 28 to 121. Duration of treatment ranged from about 3 to 5 months. Although one study did not include a followup assessment after posttreatment,⁷ the other four studies included at least one followup period 6 to 12 months posttreatment. All four studies with inactive comparators enrolled a majority of female participants;^{5, 7–9} two of these also had active comparator arms.^{5, 9} The single study that compared CT with imaginal exposure (IE)¹²⁹ had similar proportions of male and female participants. One study included those with motor vehicle accident (MVA) trauma types;⁵ the other study participants had mixed types of trauma exposures. The mean age of participants in CT studies ranged from 37 to 44. All studies used the CAPS as the primary outcome of interest.

The single MCT study that had one active comparator (PE) and one inactive comparator (wait-list) included males and females with exposure to mixed trauma types. The single CR study that contained three active comparators included males and females with exposure to mixed trauma types.¹²² Participants completed eight sessions and did not include any followup assessments after the end of treatment. The primary outcome measure was the Posttraumatic Diagnostic Scale (PDS).

Results for Cognitive Interventions Compared With Inactive Comparators

Under each outcome heading below, we first present our data synthesis for studies testing CPT against an inactive comparator. Then we present results for the CR study, CT studies, and MCT study with inactive comparator groups.

PTSD Symptoms

All included studies reported measures of PTSD symptoms. All five studies comparing CPT with inactive comparators found that subjects in the CPT arm had a greater reduction in CAPS-assessed symptoms of PTSD than those in the inactive comparator arm.^{1–4, 6} The meta-analysis that pooled CAPS scores (Figure 3) found a much greater decrease in PTSD symptoms for subjects treated with CPT therapy than for those in inactive comparator groups (SMD, −1.35; 95% CI, −1.77 to −0.94, I²=71.1%, 5 studies, N=399, moderate SOE). The meta-analysis had considerable statistical heterogeneity, but the direction of effects was consistent. The differences were only in the exact magnitude of benefit; all studies found moderate or large magnitudes of benefit. In addition, two of three studies that compared CPT with a wait-list group found that changes were maintained at 3 to 6 months (posttreatment) followup.^{2, 3}

Figure 3

Standardized mean change in PTSD symptoms (measured by the Clinician-Administered PTSD Scale) for CPT compared with inactive comparators. CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; CPT = cognitive processing therapy; PTSD = posttraumatic (more...)

All four studies that compared CT with inactive control groups reported significantly greater decreases in PTSD symptoms for those treated with CT than those in inactive comparator groups (meta-analysis not performed because of heterogeneity in sample and study characteristics, SMD of individual studies ranged from −2.0 to −0.3; 4 studies; N=236; moderate SOE).^{5, 7–9} The single study that compared MCT with an inactive comparator reported significantly greater decreases in PDS-measured PTSD symptoms, favoring MCT, as measured by the PDS (between-group mean difference=−27.7, 1 study, N=21; insufficient SOE).¹⁹

Loss of PTSD Diagnosis

Several cognitive intervention studies reported data on posttreatment diagnostic status. The four CPT studies that reported loss of diagnosis outcomes favored CPT over inactive comparator (risk difference [RD], 0.44; 95% confidence interval [CI], 0.26 to 0.62; I², 77.9%; 4 studies, N=299; moderate SOE)^1–4 (Figure 4). All four studies comparing CPT with wait-list reported followup assessments showing that, over time, the greater changes seen in loss of PTSD diagnosis among CPT participants than inactive comparator participants persisted at 1 month,¹ 12 months,^{2, 3} and 5 to 10 years after the end of treatment.¹²⁶

Figure 4

Loss of diagnosis for CPT compared with inactive comparators. CI = confidence interval; CPT = cognitive processing therapy; RD = risk difference.

Three^{5, 8, 9} of the four studies^{5, 7–9} that compared loss of PTSD diagnosis between CT and inactive comparators found significantly higher rates of loss of PTSD diagnosis at posttreatment among those who received the CT intervention as compared with those who received an inactive comparator (RD, 0.55; 95% CI, 0.28 to 0.82; I², 88.5%; 4 studies, N=314; moderate SOE) (Figure 5).

Figure 5

Loss of diagnosis for CT compared with inactive comparators. CI = confidence interval; CT = cognitive therapy; RD = risk difference.

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

All but one¹²⁷ cognitive intervention study assessed coexisting depressive symptoms using the Beck Depression Inventory (BDI) or BDI-II, and some also assessed anxiety symptoms.^1–6 Our meta-analysis of the five CPT studies reporting BDI or BDI-II scores (Figure 6) found greater improvement in depression symptoms for subjects treated with CPT than for those in inactive comparison groups (SMD, −1.09; 95% CI, −1.52 to −0.65, I²=75.0, 5 studies, N=399, moderate SOE).^{1–4, 6}

Figure 6

Standardized mean change in depressive symptoms (measured by the Beck Depression Inventory) for CPT compared with inactive controls. BDI = Beck Depression Inventory; CI = confidence interval; CPT = cognitive processing therapy; SMD = standardized mean (more...)

In three studies that included followup assessments of depressive symptoms, subjects maintained significant decreases in symptoms at 3 months^3–5 and 9 months³ after the end of treatment assessment. In another study, authors found a pre- to posttreatment depressive symptom effect size of 1.16 for CPT versus inactive comparator group, which declined to 0.49 at the 1-month followup.¹ The authors attributed the attenuation of between-group effect size differences to decreases in depressive scores in the wait-list group at followup, not to increases in depressive symptom scores in the CPT group. From the above findings and our meta-analysis, we concluded that evidence supports the efficacy of CPT for reducing depression symptoms (moderate SOE).

Two studies of CPT assessed anxiety symptoms as an outcome using the State-Trait Anxiety Inventory (STAI).^{1, 4} One found CPT to be no more effective in reducing symptoms of anxiety than wait-list;¹ the other found greater improvement in anxiety for subjects treated with CPT than those receiving usual treatment from intake to posttreatment (p=0.018).^{4, 5} We concluded insufficient evidence to determine the efficacy of CPT for reducing anxiety symptoms based on lack of consistency and imprecise findings of two studies.

All four studies that compared CT with inactive control groups assessed depressive and anxiety symptoms;^{5, 7–9} four of four studies found significant between-group differences in depression symptoms and three of four studies found significant between-group differences at posttreatment favoring the CT group (moderate SOE). In two of the studies, between-group differences favoring the CT group persisted at the 3- and 9-month followup assessments (p<0.001 for all comparisons).^{5, 7, 9}

The single MCT study reported significantly greater decreases in both depressive symptoms assessed with the BDI and anxiety symptoms assessed with the Beck Anxiety Inventory (BAI) among MCT group participants than inactive comparator participants (insufficient SOE).¹⁹

Results for Cognitive Interventions Compared With Active Comparators

Two studies tested cognitive interventions with a comparator for which we did not aim to assess comparative effectiveness (CPT versus PCT^{9, 127} and CT versus self-help booklet⁵).

Two studies compared CPT with exposure therapy,^{3, 122} one study compared CT with IE,^{129, 130} and one study compared MCT with PE.¹⁹ Assessment of these comparative effectiveness studies appears in the CBT-Exposure section below.

One study compared CPT with MEST.¹²⁴ Comparative effectiveness findings appear in the Other Psychological Interventions section below.

One study compared CR with a relaxation group and a combination of PE and CR.¹²² The CR versus relaxation comparisons appear in the CBT-Coping Skills section (below). The authors did not report data on the comparative effectiveness of CR and the combination of PE and CR.

Characteristics of Studies

Table 8 summarizes the characteristics of the six studies meeting our inclusion criteria.^{14, 46, 122, 131–133} Further details describing the included studies are provided in Appendix F.

Table 8

Characteristics of included coping skills trials.

The studies in this section had a “coping skills” arm(s)—either relaxation training, SIT, or SAT. SIT is a cognitive behavioral intervention for PTSD in which the basic goal is to help subjects gain confidence in their ability to cope with anxiety and fear stemming from trauma-related reminders. In SIT, the therapist helps patients increase their awareness of trauma-related cues for fear and anxiety. In addition, clients learn a variety of coping skills that are useful in managing anxiety, such as muscle relaxing and deep breathing. SAT contains components of SIT to include psychoeducation, skills training, and an application phase to practice new coping skills.

Two of the six studies compared coping skills interventions with inactive comparators.^{14, 46, 122, 131–133} One compared PE, SIT, combined PE+SIT, and a wait-list group,¹⁴ and the other compared relaxation, EMDR, and usual care.⁴⁶ One enrolled women who were victims of sexual or nonsexual assault,¹⁴ and the other study enrolled combat veterans with mixed trauma types.⁴⁶ Duration of treatment ranged from 6 to 9 weeks, and both had multiple followup assessments up to 9 months after the end of treatment. The primary outcome measure for one study was the PSSI;¹⁴ and the other used the CAPS as the primary outcome measure.⁴⁶

All six studies made comparisons with active psychotherapy interventions. Sample sizes ranged from 35 to 110. Duration of treatment ranged from 6 to 16 weeks. All but one study¹³² included at least one followup assessment. Two studies enrolled combat veterans;^{46, 131, 136} one enrolled victims of sexual and nonsexual assault;¹⁴ and the other three studies enrolled heterogeneous groups of subjects with a variety of index trauma types (e.g., physical assault, road accidents, nonroad accident, witnessing a trauma or homicide, sexual assault, being held hostage, bombing, combat). Mean age for subjects in the studies ranged from 35 to 48.5. Whereas three studies had over 75 percent female participants,^{14, 132, 133} two studies had samples comprising at least 75 percent males.^{46, 131} The primary outcome for nearly all studies was the CAPS; one study used the PSS-I.¹⁴

Results for Coping Skills Compared With Inactive Comparators

PTSD Symptoms

Both studies that compared a coping skills intervention with inactive comparators reported measures of PTSD symptoms (Table 9).^{14, 46} One small study found significantly greater decreases in PTSD symptoms at posttreatment among participants in the SIT versus wait-list group.¹⁴ The other study found a greater, but nonstatistically significant reduction in PTSD symptoms in the relaxation arm as compared with the treatment-as-usual arm (insufficient SOE).⁴⁶

Table 9

Results of the coping skills interventions compared with inactive controls for PTSD.

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

Two studies reported on coexisting comorbid depression symptoms and comorbid anxiety symptoms (Table 10).^{14, 46} The study that included SIT and wait-list arms found that subjects treated with SIT had greater decreases in depression symptoms than those in the wait-list group; between-group differences in pre- to posttreatment changes in anxiety symptoms did not reach statistical significance (insufficient SOE).¹⁴

Table 10

Results at the end of treatment for depression and anxiety symptoms for coping skills interventions compared with inactive controls.

The study comparing relaxation and usual care found decreases in both depression and anxiety symptoms in the relaxation group; however, the authors reported no statistically significant between-group difference on measures of anxiety and did not provide data on between-group differences for depression (insufficient SOE).⁴⁶

Results for Coping Skills Compared With Active Comparators

Of the six studies comparing a coping skills therapy with an active comparator, four included comparisons with exposure-based interventions,^{14, 122, 132, 133} two included comparisons with EMDR,^{46, 133} two included comparisons with CBT-mixed therapies,^{14, 122} one included a comparison with CR,¹²² one included a comparison with IPT,¹³² and one compared to an active control condition for which we did not aim to assess comparative effectiveness (PTSD family education [PFE]).¹³¹ For assessment of the comparisons with exposure-based therapies, see the CBT—Exposure section (below). For assessment of the comparisons with CBT-mixed therapies, see the CBT—Mixed section (below). For assessment of the comparisons with EMDR, see the EMDR section (below). For assessment of the comparisons with IPT, see the Other Psychological Interventions section (below).

Results for Coping Skills Compared With Active Comparators: Relaxation Training Versus Cognitive Restructuring

One study assessed the comparative effectiveness of four PTSD treatments; subjects in one arm received relaxation training and subjects in another arm received CR (the other two treatment arms tested CBT-exposure interventions; the CBT-exposure section details the findings from these comparisons).¹²²

Characteristics of Studies

Table 11 summarizes the characteristics of the 25 studies meeting our inclusion criteria. Types of exposure therapy tested included IE, in vivo exposure, PE (which includes both components of IE and in vivo exposure, or modified PE, mPE), virtual reality exposure (VRE), written exposure therapy (WRE), and Concurrent Treatment of PTSD and Substance Use Disorders using Prolonged Exposure (COPE). Further details are provided in Appendix F. Of the 25 included studies, 14 studies (15 articles because one study reported similar outcomes across two separate publications) compared exposure therapy with an inactive comparator: wait-list,^{3, 11–14, 16–19, 21, 122, 137} or usual care,^{10, 15, 20, 41, 138, 139} and 17 included one or more active comparators.^{3, 12–14, 16, 19, 41, 42, 122, 129, 132, 133, 138–140}

Table 11

Characteristics of included CBT-exposure trials.

These studies generally enrolled subjects with severe or extreme PTSD symptoms.^{10, 15, 19, 20, 41, 122, 133} Sample sizes ranged from 24 to 284. Four studies (one of which presented outcomes in 2 articles) assessed followup at the end of active treatment;^{19–21, 132, 137} the remainder included posttreatment followups after 3, 4½, 6, 9, or 12 months. Fourteen of the studies enrolled a heterogeneous group of subjects with a variety of index trauma types (e.g., accident, disaster, physical assault, sexual assault, witnessing death or serious injury), 4 studies enrolled a majority of subjects with sexual assault-related PTSD,^{3, 12–14} 4 enrolled subjects with combat-related PTSD,^{15, 18, 139, 141} 1 enrolled subjects with combat- or terror-related PTSD,¹⁵ 1 enrolled natural disaster victims,¹¹ and one enrolled patients involved in an MVA.¹⁷ One involved patients with both a psychotic disorder and PTSD.¹⁶ Mean age ranged from 27 to 63. Ten studies enrolled two-thirds or more female subjects. The primary outcome for the majority of studies was some version of the CAPS (CAPS, CAPS-2, or CAPS-Sx); 4 studies identified the PSS-I as the primary outcome measure;^{12, 14, 15, 142} and 1 study used the PDS.¹⁹

Results for Exposure Therapy Compared With Inactive Comparators

PTSD Symptoms

Thirteen of the 14 studies (data for 1 study reported in 2 articles) comparing various exposure therapies with an inactive comparator reported measures of PTSD symptom change; one study compared each of two exposure therapies (VRE and PE) to placebo,¹⁸ allowing 14 comparisons. All 13 studies reported outcomes in 14 publications reported greater decreases in PTSD symptoms (outcome measures included CAPS, PSS-I, and PDS) in the exposure group than in the control group.^{3, 10–21, 137} Our meta-analysis of pooled data from these studies (Figure 7) found a greater decrease in PTSD symptoms for subjects treated with exposure than for those in control groups; the effect size was very large (SMD, −1.23; 95% CI, −1.50 to −0.97, 13 studies [14 comparisons]; N=885; I-squared, 67.5%; high SOE).

Figure 7

Standardized mean change from baseline to end of treatment in PTSD symptoms (any measure) for exposure therapy compared with inactive comparator. CI = confidence interval; PTSD = posttraumatic stress disorder; PE = prolonged exposure; SMD = standardized (more...)

Our meta-analysis of the studies reporting CAPS scores found a greater decrease in PTSD symptoms among subjects treated with exposure than those in an inactive comparator group (SMD, −1.12; 95% CI, −1.42 to −0.82; 8 studies [9 comparisons], N=689; I-squared=68.66%; high SOE) (Figure 8).

Figure 8

Standardized mean change from baseline to end of treatment in PTSD symptoms (CAPS) for exposure therapy compared with inactive comparator. CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; PE = prolonged exposure; SMD = standardized (more...)

Among those studies that assessed followup measures longer-term, the effects for decreases in PTSD symptoms were maintained at 3, 6, 9, or 12 months.

Loss of PTSD Diagnosis

Six of the studies comparing subjects who received exposure therapy with those in inactive comparator groups reported loss of PTSD diagnosis between groups. Participants treated with exposure therapy had greater rates of loss of PTSD diagnosis as compared with participants in the inactive comparator groups (RD, 0.56; 95% CI, 0.35 to 0.78; 6 studies, N=409; high SOE) (Figure 9).

Figure 9

Loss of diagnosis for exposure therapy compared with inactive comparator. CI = confidence interval; RD = risk difference.

Prevention or Reduction of Comorbid Conditions

Ten studies (from 11 publications and involving 11 comparisons) with inactive comparators reported on changes in depression symptoms as measured by the BDI.^{3, 11–15, 18–21, 137} All but 1²⁰ reported a significantly greater decrease in depression symptoms for exposure intervention patients than for inactive comparators. Results of our meta-analysis indicated a greater reduction in BDI depressive symptom scores for subjects treated with exposure than for those in wait-list or usual-care inactive comparator groups (SMD, −0.76; 95% CI, −0.91 to −0.60; I²=19.4%, 10 studies [11 comparisons], N=7,152, Figure 10; high SOE).

Figure 10

Standardized mean change from baseline to end of treatment in depression symptoms (BDI) for exposure therapy compared with inactive comparator. BDI: Beck Depression Inventory; CI = confidence interval; PE = prolonged exposure; SMD = standardized mean (more...)

Three studies reported on anxiety symptoms, two using the Beck Anxiety Inventory^{18, 19} and one using the State Trait Anxiety Inventory.²⁰ The two using the Beck Anxiety Inventory indicated significant benefit for exposure therapy, while the study using the State Trait Anxiety Inventory (which involved patients with comorbid substance dependence) did not show significant benefit (low SOE).

Results for Exposure Therapy Compared With Active Comparators

The 17 studies that compared exposure therapy with an active comparator included comparisons with EMDR,^{13, 16, 133} a coping skills intervention (relaxation training ^{122, 132, 133} or SIT¹⁴), a cognitive intervention (CPT, CR, CT, or MCT),^{3, 19, 122, 129} IPT,¹³² PE+CR,^{12, 42, 122} IE+CR,⁴¹ PE plus SIT,¹⁴ IRT¹⁴² or another type of active control (PCT,^{138, 139} relapse prevention,¹⁴³ active monitoring comparison group [AMCG],¹⁴³ supportive counseling [SC],⁴¹ or a health information–based active control).¹⁴⁰

In this section, we address the 17 studies comparing CBT-exposure therapy with an active comparator. We do not report on comparisons between exposure therapy arms and interventions for which we did not aim to assess comparative effectiveness (i.e., comparisons between different types of exposure interventions^{39, 40, 144–148} or between exposure interventions and PCT, SC, relapse prevention, AMCG, or health information comparators).^{3, 12–14, 16, 18, 19, 41, 42, 122, 129, 132, 133, 140, 143}

Results for Exposure Therapy Compared With Active Comparators: Exposure Therapy Versus Cognitive Interventions

Four studies compared exposure therapy and either CPT, CR, CT, or MCT.^{3, 19, 122, 129} Of these, one compared PE with CR,¹²² one compared IE with CT,¹²⁹ one compared PE with CPT,³ and one compared PE with MCT.¹⁹ We did not perform quantitative meta-analysis to pool the findings because each intervention and comparator were different across each of the 4 studies.

Results for Exposure Therapy Compared With Active Comparators: Exposure Therapy Versus Coping Skills Therapies

Four studies compared exposure therapy with a coping skills therapy.^{14, 122, 132, 133} One compared PE with SIT,¹⁴ and the others compared PE with relaxation therapy.

Results for Exposure Therapy Compared With Active Comparators: Exposure Therapy Compared With Eye Movement Desensitization and Reprocessing

Results for Exposure Therapy Compared With Active Comparators: Exposure Therapy Versus Exposure Plus Cognitive Restructuring

Four studies compared exposure therapy with exposure+CR.^{12, 41, 42, 122} Three tested PE against PE+CR,^{12, 42, 122} whereas the other tested IE against IE+CR.⁴¹

Results for Exposure Therapy Compared With Active Comparators: Prolonged Exposure Versus Interpersonal Psychotherapy

As noted above, one study compared PE (N=38), IPT (N=40), and relaxation (N=32).¹³² We previously compared exposure and relaxation; here, we report the PE versus IPT comparison (insufficient SOE for each outcome). Both types of interventions led to substantial decreases in PTSD symptoms at posttreatment, but the authors found no significant between-group differences. In addition, the proportions of subjects who entered remission did not differ (RD=0.03); the groups also did not differ with respect to decreases in depressive symptoms as measured by the HAM-D or changes in quality-of-life ratings as measured by the Quality of Life Enjoyment and Satisfaction scale.¹³² We concluded that evidence is insufficient to determine the comparative effectiveness of PE versus IPT for PTSD symptoms, remission, depressive symptoms, and quality of life based on this single study.

Results for Exposure Therapy Compared With Active Comparators: Prolonged Exposure Versus Imagery Rehearsal Therapy

One study compared PE and IRT.¹⁴² Significant differences were not found for decreases in PTSD symptoms, percentage recovering or with symptom improvement, or decreases in depression symptoms or psychological symptom severity between treatment groups. The treatment* time interaction for the psychological subscale of the quality of life assessment scale approached significance (p=0.05), with the PE group participants having slightly greater improvements at posttreatment that evened out at the 3-month followup assessment. We concluded that evidence is insufficient to determine the comparative effectiveness of PE versus IRT for PTSD symptoms, remission, depressive symptoms, and quality of life based on this single study.

Characteristics of Studies

Table 12 summarizes the characteristics of the 31 studies meeting our inclusion criteria. Further details about these studies appear in Appendix D. The studies in this section are somewhat heterogeneous in several ways: how authors define and describe “cognitive behavioral therapy,” duration of the intervention, and mode of delivery. Elements of the CBT arm of the studies considered here include psychoeducation, self-monitoring, stress management, relaxation training, skills training, exposure (imaginal, in vivo, or both), cognitive restructuring, guided imagery, mindfulness training, breathing retraining, crisis/safety planning, and relapse prevention. The studies varied as to how many sessions (if any) were dedicated to these elements and whether homework was assigned as part of the intervention.

Table 12

Characteristics of included CBT-mixed intervention trials.

Twenty-three of these 31 studies included an inactive comparator, such as a wait-list (16 studies), usual care (4 studies), or SC (4 studies).^{12, 14, 22–41, 149} Thirteen of the 31 studies made comparisons with active interventions (i.e., other psychotherapies).^{12, 14, 39–42, 122, 144–146, 148, 151, 152} Of these 13 studies, 5 included an exposure-based intervention as the comparison;^{12, 14, 41, 42, 122} 1 used structured writing therapy (SWT);⁴⁰ 1 used PCT;³⁹ 2 used relaxation;^{122, 152} 2 used another CBT-mixed intervention;^{148, 151} 1 used a CBT for Alcohol Use Disorder (AUD) plus SC;¹⁴⁶ 1 used a group integrated cognitive behavioral therapy (ICBT) for depression and substance use disorder (SUD) followed by individual ICBT for depression and SUD;¹⁴⁵ and 1 study that tested a dialectical behavior therapy-(DBT)-PE combination therapy used DBT alone as the comparison group.¹⁴⁴

Of the 24 studies with inactive comparators, sample sizes ranged from 23 to 190. Duration of treatment ranged from 4 to 24 weeks. Although 3 studies did not include a followup assessment^{22, 24, 25} and the followup interval for 1 was unclear,⁴⁰ the remainder of the studies with inactive comparators included at least one posttreatment followup assessments after 1 to 12 months.⁴⁰ The majority of studies enrolled a heterogeneous group of subjects with a variety of index trauma types and comorbid mental health problems (e.g., depression, personality disorders, SUD/AUD). Mean age ranged from 30 to 61 years. Most studies enrolled a large majority of female subjects. The primary outcome measure for 13 of these studies was some version of the CAPS (CAPS, CAPS-2, or CAPS-Sx);^{22–24, 27–29, 34–39, 41} 4 studies used a form of the PSS (PSS-I or PSS-SR);^{12, 14, 32, 33} 3 studies used the PDS;^{23, 24} 5 studies used the PCL;^{25, 26, 30, 31, 149, 153} and 2 the IES.^{24, 40}

Of the 13 studies with active comparators,^{40–42, 122} sample sizes ranged from 24 to 190. Duration of treatment ranged from 8 to 16 weeks, with the exception of the study testing DBTPE vs. DBT,¹⁴⁴ where treatment lasted for 1 year. All studies also included posttreatment followup assessments ranging from 1 to 12 months. The majority of studies enrolled a heterogeneous group of subjects with a variety of index trauma types and comorbid mental health conditions (e.g., depression, SUD/AUD, personality disorders, intentional self-injury). Mean age ranged from 33 to 50. Most studies enrolled a large majority of female subjects. The primary outcome for 6 studies was some version of the CAPS (CAPS, CAPS-2, or CAPS-Sx); 3 used the PSS-I,^{12, 14, 146} 2 used the PCL,^{144, 152} 1 used the PDS,¹⁴⁶ and 1 used the IES.⁴⁰

Results for CBT-Mixed Interventions Compared With Inactive Comparators

PTSD Symptoms

Of the 23 studies with inactive comparators, 21 compared PTSD symptom changes pre- to posttreatment between CBT-M and inactive comparator groups, half (n=11) of which used the CAPS to report outcomes. Among the 11 studies that used the CAPS, 8 reported decreases in PTSD symptoms as assessed by the CAPS that were statistically significant. Full evidence tables from each of these studies can be found in Appendix F.

Our meta-analysis (Figure 11) found greater decreases in CAPS-rated PTSD symptoms for CBT-M interventions than for inactive controls (SMD, −1.24; 95% CI, −1.67 to −0.81; 11 studies, N=709; high SOE). Statistical heterogeneity was substantial (I²=85.4%). Much of the heterogeneity may be explained by the diversity of both interventions (as explained above, these interventions used various CBT components). Six studies found a similarly large decrease in PTSD scores assessed by CAPS for CBT-M intervention groups compared with wait-list or usual-care controls—about a 30-point greater reduction.^{23, 28, 34, 36–38} One study with a wait-list control found even greater decreases (about a 68-point decrease).³⁵ Three of the 10 studies found little to no decrease.^{27, 29, 39} One of these compared CBT-M interventions with usual care (in which the control patients were often receiving some form of treatment)²⁹ and another with standard care²⁷ rather than with wait-list; this likely biased results toward the null.

Figure 11

Mean change from baseline in CAPS for CBT-mixed interventions compared with inactive comparators. CAPS = Clinician-Administered PTSD Scale; CBT = cognitive behavioral therapy; CI = confidence interval SMD = standardized mean difference.

We conducted additional meta-analyses to pool pre- to posttreatment differences in PTSD symptoms across groups using additional outcome measures reported across all studies with inactive comparators (CAPS, PSS-I, IES, PCL, PDS). Our meta-analysis found greater decreases in PTSD symptoms for CBT-mixed interventions compared with inactive controls (SMD, −1.01; 95% CI, −1.28 to −0.74; 21 studies, N=1,349, Appendix H; high SOE). Similar to the CAPS meta-analysis, statistical heterogeneity was substantial (I²=81.1%). However, also like the synthesis of CAPS data, the differences in findings were in the magnitude (not the direction) of the effect; all point estimates favored CBT-mixed interventions, and the vast majority of individual studies reached statistical significance.

Three of the 10 studies reported data on PTSD symptoms assessed by CAPS at 3- to 6-month followups (Appendix H).^{23, 29, 36} Of these, 2 found significant between-group differences favoring the CBT-mixed intervention over inactive comparators,^{23, 36} and the other study found no significant differences between groups.²⁹ These findings mirrored those found at the end of treatment assessments.

Adding three additional studies that used PTSD symptom measures other than the CAPS to the analysis permitted pooled analysis via meta-analysis. Of the six studies, four found statistically significant between-group differences from pretreatment to followup assessment, favoring CBT-mixed interventions over wait-list^{23, 32, 36} and SC groups;³³ two studies found no significant pretreatment to followup assessment between a CBT-mixed intervention and wait-list²⁶ or usual care (Appendix H).²⁹ Meta-analysis of the six studies found that between-group differences in PTSD symptom changes persisted after the end of treatment but with a somewhat smaller effect size than found at the posttreatment assessment (SMD, −0.8; 95% CI, −1.3 to −0.2; Appendix H; high SOE). Determining with confidence how much of the between-group differences in PTSD symptom decreases persist at longer-term followup is difficult, partly because of the potential for reporting bias (i.e., studies not reporting followup data because the significant differences did not persist).

Loss of PTSD Diagnosis

Nine studies reported sufficient data on loss of PTSD diagnosis to permit meta-analysis.^{22–24, 31–34, 39, 41} Our meta-analysis (Figure 12) found a large effect size (RD, 0.29; 95% CI, 0.17 to 0.40; I²=58.1%, 9 studies, N=474) for loss of PTSD diagnosis between CBT-mixed and inactive comparator subjects (high SOE).

Figure 12

Loss of PTSD diagnosis for CBT-mixed interventions compared with inactive comparator. CBT = cognitive behavioral therapy; CI = confidence interval; RD = risk difference.

Two of the studies also reported 3- to 6-month loss of PTSD diagnosis followup data.^{33, 41} Significant findings from both studies suggested that the between-group differences in loss of PTSD diagnosis favoring the CBT-mixed interventions over inactive comparators persisted over time.

Disability or Functional Impairment

Six studies reported data on disability or functional impairment^{23, 30–32, 36, 37} using a variety of measures (Table 13). We did not use meta-analysis to pool findings because of the diversity of measures that did not assess the same types of disability and functional impairment. Four of the six studies found significantly greater improvements in disability or functional outcomes for those who received CBT-mixed interventions than those who received an inactive control (low SOE).

Table 13

Results at the end of treatment for disability or functional impairment outcomes for CBT-mixed interventions compared with inactive controls.

Results for CBT-Mixed Interventions Compared With Active Comparators

Of the 12 studies comparing a CBT-mixed intervention with an active comparator, 5 compared a CBT-mixed intervention against an exposure-based intervention.^{12, 14, 41, 42, 122} Assessment of head-to-head comparisons with exposure-based interventions is covered in the CBT—Exposure section (above). Several of the other studies made comparisons with interventions for which we did not aim to assess comparative effectiveness^{39, 40, 144–146, 148} (e.g., comparisons with other CBT-mixed interventions,^{144–146, 148} SWT,⁴⁰ or an active control [PCT].³⁹ In this section, we address the 2 studies comparing CBT-mixed interventions and relaxation interventions.^{122, 152}

Results for CBT-Mixed Interventions Compared With Active Comparators: CBT-Mixed Versus Relaxation

Characteristics of Studies

Table 14 summarizes the characteristics of the 10 studies meeting our inclusion criteria. Further details describing the included studies are provided in Appendix F. Eight studies had an inactive comparator, such as wait-list,^{13, 16, 44–46, 48} “stabilization as usual,”⁴³ or placebo.⁴⁷ Five had active comparisons with either PE,^{13, 16, 133} BEP,¹⁵⁴ or relaxation training.^{46, 133}

Table 14

Characteristics of included EMDR trials.

Sample sizes ranged from 21 to 155. Duration of treatment ranged from 4 to 8 weeks. Although one study did not include a followup assessment⁴⁸ and another included only a followup at 5 weeks posttreatment, the rest of the EMDR studies included posttreatment followups after 3, 6, or 9 months. Two of the studies enrolled a heterogeneous group of subjects with a variety of index trauma types (e.g., sexual assault, physical assault, witnessing traumatic events, accidents, and combat), one study enrolled a majority of subjects with combat-related PTSD,⁴⁶ one enrolled Swedish public transportation workers who witnessed train accidents or were physically assaulted,⁴⁸ two enrolled female victims of sexual assault,^{13, 45} two enrolled refugees,^{43, 44} and one enrolled participants with comorbid psychotic disorders with mixed trauma types.¹⁶ Mean age was roughly similar across studies, ranging from 34 to 49 years. Four studies enrolled 70 percent or more female subjects.^{13, 44, 45, 47} The primary outcome for the majority of studies was some version of the CAPS (CAPS, CAPS-2, or CAPS-Sx); two studies identified other primary outcomes, including the PSS-I⁴⁵ or IES.^{44, 48}

Among the studies with inactive comparators described above, two also included an active comparator arm of either PE¹³ or relaxation.⁴⁶ Another study compared EMDR with either PE or relaxation therapy in a sample of individuals with PTSD with mixed trauma exposure types.¹³³ A fourth study included an active comparator (PE) and an inactive comparator (wait-list).¹⁶

Results for EMDR Compared With Inactive Comparators

PTSD Symptoms

All eight EMDR studies with inactive comparators measured PTSD symptom change. Our meta-analysis (Figure 13) found greater decreases in PTSD symptoms for EMDR than for inactive comparator subjects (SMD, −1.08; 95% CI, −1.82 to −0.35; I squared=91.5%, 8 studies; N=449).^{13, 16, 43–48} Differences between EMDR and comparator groups reached statistical significance in four of eight studies;^{13, 16, 44, 45} point estimates varied widely across studies (moderate SOE). The two studies that found a significant pre- to posttreatment benefit of EMDR and included followup assessments reported the maintenance of benefit at the 1-month⁴⁴ and 9-month¹⁶ followup assessments.

Figure 13

Standardized mean change from baseline in PTSD symptoms for EMDR compared with inactive comparator. CI = confidence interval; EMDR = eye movement desensitization and reprocessing; PTSD = posttraumatic stress disorder; SMD = standardized mean difference. (more...)

Loss of PTSD Diagnosis

Of the studies that compared EMDR with inactive comparators, seven of the eight studies reported sufficient data to permit meta-analysis. Although one study that compared EMDR to stabilization as usual⁴³ did not find significant differences between groups, the other seven studies found a greater loss of diagnosis among EMDR subjects than inactive comparator subjects at posttreatment and at followup assessments.^{13, 16, 44, 45, 47, 48} Our meta-analysis (Figure 14) found large between-group differences (RD, 0.43; 95% CI, 0.25 to 0.61) in loss of diagnosis at posttreatment assessment (moderate SOE).⁴⁷

Figure 14

Loss of PTSD diagnosis for EMDR compared with inactive comparators.

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

Seven studies comparing EMDR with inactive comparators included a measure of depression symptoms (BDI, HAM-D, or Hopkins Symptom Check List [HSCL]-depression). All demonstrated better improvements in depression symptoms among EMDR group subjects, with four of the seven study comparisons reaching statistical significance.^{13, 43, 44, 46, 48} Our meta-analysis (Figure 15) indicated a significant effect size (SMD, −0.91; 95% CI, −1.58 to −0.24; 7 studies; I squared=87.5%, N=347; moderate SOE).

Figure 15

Standardized mean change from baseline in depression symptoms for EMDR compared with control. CI = confidence interval; EMDR = eye movement desensitization and reprocessing; SMD = standardized mean difference.

Three studies used STAI,^{13, 45, 46} and one used the anxiety subscale of the HSCL⁴³ to assess anxiety symptoms. Although all studies found that EMDR improved anxiety symptoms more than inactive controls, results did not reach statistical significance in three of the four studies (meta-analysis not conducted because of heterogeneity in sample characteristics, insufficient SOE).^{13, 43, 45, 46}

Quality of Life

One study assessed quality-of-life outcomes.⁴³ Differences in quality-of-life changes from pre- to posttreatment did not significantly differ between the EMDR and stabilization-as-usual groups (insufficient SOE).

Results for EMDR Compared With Active Comparators: Relaxation

Of the studies comparing EMDR with an active comparator, three compared EMDR and exposure therapy;^{13, 16, 133} as assessed in the CBT—Exposure section (above); one study compared EMDR with BEP¹⁵⁴ as assessed in the Other Psychological Intervention section (below). Two studies compared EMDR and relaxation.^{46, 133}

Characteristics of Studies

Table 15 summarizes the characteristics of 24 studies meeting our inclusion criteria. Further details describing the included studies are provided in Appendix D. For the characteristics and results sections included in this section, we group studies primarily by intervention type (rather than comparator) because of the large heterogeneity across studies.

Table 15

Characteristics of included studies of other psychological interventions.

Two studies assessed TAR (full intervention name: Trauma Affect Regulation: Guide for Education and Therapy). One study compared individually delivered TAR with two comparison groups (PCT and wait-list) in a population of mothers with victimization-related PTSD.⁵⁹ The second study compared TAR delivered in a group setting with supportive group therapy in a population of incarcerated women victims of interpersonal violence.⁶⁰ Enrolled populations had a mean age of 31 to 36 years and had about half nonwhite participants (range 43% to 59%).

Four studies assessed BEP. Three of the four BEP studies had wait-list^{50, 51} or minimal attention comparators;⁴⁹ one compared BEP with EMDR.¹⁵⁴ Three studies conducted by the same research group in the Netherlands had varying sample characteristics; one sampled police officers,⁵¹ and the other two had heterogeneous subjects with a variety of index trauma types.^{50, 154} Treatment lasted for 16 weeks in all four BEP studies, with similar mean age of participants across studies (35 to 40 years of age). Twelve subjects (40%) of the Swiss sample were taking psychotropic medications, “mostly antidepressants.”

Three studies evaluated IRT.^{52, 142, 156} One IRT study that tested efficacy versus wait-list involved women with a history of sexual trauma (N=168).⁵² Another study compared IRT with psychoeducation in male Vietnam-era combat veterans with no medical disorders known to affect sleep (e.g., narcolepsy, untreated sleep apnea).¹⁵⁶ A third study tested the comparative effectiveness of IRT with PE among men and women with mixed trauma types. All studies included a 3-month followup posttreatment.

Four studies assessed the effectiveness of NET for PTSD among asylum seekers and refugees. Sample sizes ranged from 32 to 277. Duration of treatment ranged from 3 to 12 weeks. Three studies used the PDS to assess PTSD symptom severity, and one used the CAPS.⁵⁵ All samples contained males (25% to 69%) and females (31% to 75%) with mean ages ranging from 28 to 35 years. One study compared NET (n=17), SC (n=14), and psychoeducation (n=12) in a Ugandan refugee settlement with Sudanese refugees.¹⁶¹ The second study, also conducted in a Ugandan refugee settlement, compared NET (n=111), trauma counseling (n=111), and a nontreatment symptom monitoring group (n=55) among Rwandan and Somalian refugees.⁵⁴ The third study compared NET (n=16) with treatment as usual (n=16) in a sample of asylum seekers living in Germany who were originally from Turkey, the Balkans, or Africa.⁵³ The fourth study compared NET (n=17) with a wait-list control (n=17) in a sample of refugees and asylum seekers from Africa and the Middle East who were living in Switzerland.⁵⁵

Two studies tested MEST. One enrolled Iranian combat male veterans and compared outcomes against those who received a control treatment¹⁶⁰ while the other tested the comparative effectiveness of MEST and CPT among men and women with mixed trauma types.¹²⁴ Both included followup assessments 3 months postintervention.

Two studies tested MBSR in samples of male and female war veterans. One tested MBSR plus treatment as usual (TAU) versus TAU,¹⁵⁹ and the other tested MBSR versus an active control, group PCT.¹³⁶ MBSR is a treatment that uses meditation to increase awareness of present mental and physical processes. In MBSR, the instructor leads participants through meditative exercises that focus on noticing sensations, thoughts, and emotions without judgment, and the participants practice short guided meditation exercises outside of group sessions. Three studies assessed other (unique) psychological interventions including the following: IPT,¹³² EFT,¹⁵⁵ and neurofeedback training.¹⁶² Appendix A details characteristics of each of these treatments. The study assessing IPT had two active comparator groups: PE and relaxation therapy.¹³² The other two studies compared an intervention with an inactive control. The enrolled population in each study had different trauma types (Table 15).

Four studies assessed the efficacy of SS; three different active control approaches contained components to treat SUDs alone or to provide psychoeducation about women’s health issues.^{56, 57, 157} One of these three studies compared the addition (to TAU) of a voluntary SS intervention with a treatment-as-usual control group, which comprised incarcerated women enrolled in a residential substance use treatment program in a minimum security wing.⁵⁶ Another active control involved TAU in a SUD clinic at a Veteran’s Administration outpatient mental health clinic.⁵⁸ Three of the studies enrolled women generally in their 30s; one enrolled male veterans with a mean age of 54.⁵⁸ Sample sizes ranged from 49 to 353;^{56, 57, 157} one of these was a pilot study (N=49) that may have been underpowered.⁵⁶ One study enrolled a sample of incarcerated women;⁵⁶ two enrolled community-based samples of women seeking substance abuse treatment.^{57, 157} Followup assessments were conducted at 3 and 6 months in all studies; one study each conducted additional assessments at 9 months⁵⁷ or 12 months.¹⁵⁷

PTSD Symptoms

Two studies assessing TAR in populations of women with interpersonal victimization reported between-group changes in CAPS scores (low SOE). The study that compared TAR, PCT, and wait-list reported greater improvement in CAPS-assessed PTSD symptoms for those treated with TAR than those in the wait-list group (between-group mean difference, −17.4; p<0.001).⁵⁹ The study that compared TAR with usual group care for incarcerated women found a similar reduction in CAPS-assessed PTSD symptoms between groups (between-group mean difference, −2.7; p=NS).⁶⁰

Remission

Both studies reported on the percentage of participants with full remission from baseline to followup with inconsistent findings. One study found a higher rate of remission among the TAR group than among the wait-list group (RD, 0.21; p<0.001);⁵⁹ the other found a lower rate of remission among the TAR group than among the usual-care group (RD, −0.11; insufficient SOE).⁶⁰

Loss of Diagnosis

Both studies reported on loss of PTSD diagnosis from baseline to followup; one found a higher rate of remission in the TAR group than in the wait-list group (RD, 0.26), and the other found a similar rate of remission in both groups (RD, 0.01; p=NR; insufficient SOE).⁶⁰

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

The study that compared TAR with a wait-list control reported greater decreases in depression symptoms and anxiety symptoms for the TAR than for the wait-list group (BDI between-group mean difference, −4.1; p<0.01; STAI between-group mean difference, −6.3; p=0.19; insufficient SOE).

PTSD Symptoms

Two studies reported measures of PTSD symptoms for BEP compared with an inactive comparator.^{49, 50} One only reported subscale scores, however, for the Structured Interview for PTSD (SI-PTSD) measure between groups.⁵⁰ The study that used CAPS reported significantly greater decreases in PTSD symptoms in the BEP versus the wait-list group (between-group mean difference=−10.8; 1 study; N=30; insufficient SOE).^{49, 51}

The study that compared PTSD symptoms between BEP and EMDR groups reported that both treatments were equally effective in reducing PTSD symptom severity, but EMDR resulted in faster recovery.¹⁵⁴ The study reported significant decreases in PTSD symptoms within both treatment groups using the IES-R and the SI-PTSD but greater decreases from baseline to the first assessment for those treated with EMDR than for those treated with BEP (between-group mean difference on SI-PTSD −10.80; 95% CI, −15.23 to −6.37).¹⁵⁴ The between-group difference did not remain significant at the second assessment, conducted after both groups had completed treatment (insufficient SOE).

Remission

One study (N=30) reported data on symptom remission. At the end of treatment, a greater proportion of BEP than inactive comparator group subjects had remitted (RD 0.13) as defined by having a CAPS score of less than 20.⁴⁹ Difference persisted at the 6-month followup (RD 0.19). None of the subjects in the wait-list group achieved complete remission at either assessment (insufficient SOE).

Loss of PTSD Diagnosis

All three BEP studies reported efficacy for loss of PTSD diagnosis at the end of treatment and followup assessments using different assessment measures. The RDs ranged from 0.13 to 0.58 in individual studies.^49–51 We concluded that evidence supports the efficacy of BEP for loss of PTSD diagnosis (low SOE).

The study that compared BEP with EMDR reported similar rates at the end of treatment (RD=0.08 slightly favoring EMDR), but EMDR subjects had quicker time to loss of PTSD diagnosis than BEP subjects (RD at mid-treatment 0.40 favoring EMDR; insufficient SOE).¹⁵⁴

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

All three studies comparing BEP with wait-list reported on reduction of depression and anxiety symptoms. Two used the Hospital Anxiety and Depression Scale (HADS) as an outcome measure; both studies reported that BEP subjects had significantly greater decreases in depression symptoms at the end of treatment and followup than wait-list subjects and at later followup (low SOE).^{49, 50} One study used the Symptom Checklist-90 (SCL-90) as a multidimensional indicator of psychopathology and reported that BEP had greater decreases in depression symptoms than wait-list at the end of treatment (data not reported, p<0.01) that persisted at the 3-month followup.⁵¹

Two studies reported that BEP had significantly greater decreases in anxiety symptoms (low SOE) as assessed by the HADS (Cohen’s d=0.8, p<0.05 and d=0.9, p<0.05 for one study at the end of treatment and at followup;⁴⁹ for the other study d=0.54⁵⁰). The study using the SCL-90 reported that BEP had greater decreases in anxiety symptoms at the end of treatment and at the 3-month followup (data not reported, p-values of <0.05 and <0.01).⁵¹

The study comparing BEP with EMDR reported measures of depression and anxiety symptoms (using the HADS depression and the HADS anxiety).¹⁵⁴ Similar to findings for other outcomes (e.g., PTSD symptoms), the study reported greater improvement from baseline to the first assessment for those treated with EMDR than for those treated with BEP but no significant difference between groups at the second assessment (insufficient SOE, see Appendix F for detailed data).

Return to Work or Active Duty

Two studies reported outcomes related to work (insufficient SOE)—one reported the percentage of subjects on sick leave;⁵⁰ the other reported the percentage who had returned to work.⁵¹ The former study found fewer subjects on sick leave for the BEP group compared with those on the wait-list, but the difference was not statistically significant (d=0.33, p=0.06).⁵⁰ The second study reported significantly greater rates of returning to work among BEP subjects than those in the inactive comparator group (RD 0.26; p<0.05) for return to work.⁵¹

PTSD Symptoms

One IRT study found efficacy for decreases in CAPS-assessed PTSD symptoms among IRT-treated versus wait-list control groups (between-group mean difference, −21.0; p=0.001; low SOE).⁵²

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

One study that assessed the efficacy of IRT on depressive symptoms using the HAM-D found greater, but not significantly different, decreases in depression among IRT participants compared with wait-list participants (insufficient SOE).⁵² This study also found significant (p<0.04) differences in changes in anxiety symptoms between groups at the end of treatment, but differences resulted from the IRT group having decreases in symptoms and the wait-list group having increases in symptoms.

PTSD Symptoms

All three studies that compared NET with an inactive comparator found that NET subjects had greater decreases in PTSD symptoms at the end of treatment (moderate SOE; Figure 16).^53–55 One study reported a reduction (but no data) in PTSD symptoms for subjects in the intervention group at the followup assessment 6 months after the end of treatment;⁵³ another reported that the intervention led to significantly greater decreases in PTSD symptoms than no treatment (i.e., monitoring group) from baseline to the 6-month followup (d=1.4 and 0.08, respectively, p<0.001).⁵⁴

Figure 16

Standardized mean change from baseline to end of treatment in PTSD symptoms for narrative exposure therapy compared with inactive controls. CI = confidence interval; NET = narrative exposure therapy; PDS = Posttraumatic Diagnostic Scale; SMD = standardized (more...)

Loss of PTSD Diagnosis

Two studies of NET and an inactive control reported data on loss of PTSD diagnosis (low SOE; Figure 17).^{53, 54} One of these studies also had an active comparator group (trauma counseling), for which we did not intend to assess comparative effectiveness.⁵⁴ Both studies with inactive comparators favored NET for loss of PTSD diagnosis at the end of treatment (RD of 0.06 and 0.14 in the two studies).^{54, 161}

Figure 17

Loss of PTSD diagnosis for narrative exposure therapy compared with inactive controls. CI = confidence interval; NET = narrative exposure therapy; RD = risk difference.

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

Two studies evaluated the efficacy of NET on coexisting psychiatric conditions.^{53, 55} Both studies that compared NET with an inactive control reported greater decreases in depressive symptoms for NET subjects compared with inactive comparator subjects at the end of treatment; however, only one comparison likely attained statistical significance (Cohens d=0.54 but p=NR; insufficient SOE).^{53, 55, 161}

One small study (N=32) found greater decreases on Composite International Diagnostic Interview-Pain scores for NET versus inactive comparator subjects at the end of treatment (insufficient SOE).⁵³

PTSD Symptoms

Participants randomized to IPT therapy had a similar decrease in PTSD symptom scores as assessed by the CAPS compared with the relaxation therapy group (between-group mean difference, −6.3 favoring IPT; p=0.097).¹³² On the PSS, however, participants in the IPT group had greater decreases in scores than the relaxation group (between-group mean difference, −18.2 favoring IPT; p=0.008; insufficient SOE).¹³²

Remission

The study that compared IPT versus relaxation found similar rates of PTSD remission (defined as CAPS score <20) across groups (RD 0.04 favoring IPT; insufficient SOE).¹³²

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

The comparative effectiveness of IPT and relaxation for change in depression symptoms as assessed by the HAM-D did not differ at the end of treatment (between-group mean difference, 0.3; insufficient SOE).¹³²

Quality of Life

IPT subjects had greater increases in quality-of-life ratings as assessed by the Quality of Life Enjoyment and Satisfaction Questionnaire¹⁶³ than the relaxation therapy group (between-group mean difference, 10.1 favoring IPT; p<0.001; insufficient SOE).¹³²

Functional Impairment

IPT subjects had greater increases in interpersonal functioning as measured by the Inventory of Interpersonal Problems than those in the relaxation group (between-group mean difference, −0.46 favoring IPT; p=0.001; insufficient SOE).¹³²

PTSD Symptoms

One study enrolling Iranian combat veterans compared MEST with a no treatment control group; participants in the MEST group had significantly fewer PTSD symptoms on the IES-R than the control group at followup (insufficient SOE).¹⁶⁰

The MEST comparative effectiveness study found that reductions in PTSD symptoms were not similar and not statistically different among subjects randomized to the MEST group and the CPT group. The MEST group participants experienced similar decreases in symptoms after attending only about half of the sessions attended by the CPT group participants (insufficient SOE).¹²⁴

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

The study reported no significant difference between participants in the MEST group and controls¹⁶⁰ or MEST versus CPT¹²⁴ on depression symptoms measured by the BDI-II (scores not reported; insufficient SOE for both efficacy and comparative effectiveness).

PTSD Symptoms

No significant differences were found between reduction in PTSD symptoms at posttreatment between those in the MBSR+TAU group versus the TAU group (insufficient SOE).¹⁵⁹

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

Differences in posttreatment decreases in depressive symptoms as measured by the Patient Health Questionnaire (PHQ-9) were not significantly significant across treatment groups (MBSR+TAU versus TAU; insufficient SOE).¹³²

PTSD Symptoms

The NF group had significantly greater decreases in PTSD as measured by the CAPS (between-group mean difference, −26.8, p<0.05) and by the Davidson Trauma Scale (DTS) (mean=−20.7, p<0.05) than the wait-list group at the posttreatment assessment (insufficient SOE).¹⁶² Between-group differences were sustained at the 1-month followup assessment.

Loss of PTSD Diagnosis

In the single small study that met review inclusion criteria, loss of diagnosis comparisons favored the NF group over the wait-list group (RD, 0.40, p=0.0002; insufficient SOE).¹⁶²

PTSD Symptoms

One study enrolling U.S. veterans compared EFT with wait-list control; participants in the EFT group had a greater decrease in PTSD symptoms measured by the PCL-M than controls at the end of treatment assessment (between-group mean difference, −22.1; p<0.0001; insufficient SOE).¹⁵⁵

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

A single study found significantly greater decreases in psychiatric symptoms at the end of treatment for EFT than inactive comparator subjects for both domains tested from the Symptom Assessment-45 Questionnaire (insufficient SOE).¹⁵⁵

PTSD Symptoms

Four studies tested the efficacy or effectiveness of SS.^{56–58, 157} Three compared SS with usual care,^56–58 and two compared SS with an active comparator, but not ones for which we sought to determine comparative effectiveness (i.e., psychoeducation,¹⁵⁷ relapse prevention⁵⁷).

The three studies comparing SS with usual care each found that the intervention participants had greater decreases in PTSD symptoms than usual-care participants; however, between-group differences did not reach statistical significance (meta-analysis not performed because of heterogeneity in sample and study characteristics, low SOE for no difference).^56–58 Figure 18 shows the SMD and CIs for between-group differences in PTSD symptoms.

Figure 18

Mean change from baseline in PTSD symptoms for Seeking Safety compared with inactive comparator. CI = confidence interval; PTSD = posttraumatic stress disorder; SMD = standardized mean difference; SS = Seeking Safety.

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

Analyses of data from two⁵⁷ studies compared substance use outcomes between SS and usual-care subjects. At the end of treatment, one study found no between-group differences for abstinence and substance use severity at the end of treatment,⁵⁶ and the other study found significantly greater decreases in drug use (but not alcohol use) among SS subjects compared with usual-care subjects at the end of treatment assessment (insufficient SOE).⁵⁸

Characteristics of Included Studies

Table 16 summarizes the characteristics of the five included studies (2 of which examined different moderators of the same trial). Each study included a subgroup analyses of trials that have been described in previous parts of this report. Each examined a different treatment comparison, although three studies (2 of which examined different moderators of the same trial) included PE as one of the treatment arms.^{3, 132, 135} Only one of the four studies (2 moderator studies from the single trial) did not include a followup after posttreatment assessment.^{132, 135} All used the CAPS as the primary outcome measure. Additional details describing the included studies can be found in Appendix F.

Table 16

Characteristics of included studies that compared efficacy or comparative effectiveness between patients having different characteristics or specific trauma types.

One study examined the efficacy and comparative effectiveness of CPT and PE (versus a wait-list comparator) within a sample of female rape survivors enrolled in a clinical trial (n=171)³ who also had childhood sexual abuse (n=121).¹²⁵ The second study compared EMDR, fluoxetine, and placebo in subjects with a variety of trauma types including child sexual abuse, child physical abuse, child sexual and physical abuse, adult sexual assault, adult physical assault, domestic violence, other adult victimization, traumatic loss, war/terror/violence, and injury/accident.⁴⁷ The authors reported subgroup analyses for those with child-onset trauma and those with adult-onset trauma. The third study that included subgroup comparisons tested the efficacy of DBT against a wait-list, treatment-as-usual group of child abuse survivors.²³ The efficacy of DBT was compared between those with and without borderline personality disorder. The fourth and fifth studies, which tested moderators from the same trial, examined the comparative effectiveness of PE, IPT, and relaxation training among a group of adults with chronic PTSD of mixed trauma types.^{132, 135} Subgroup analyses compared the effectiveness of different treatments among adults with and without major depressive disorder, trauma type, gender, and age of primary trauma exposure.

Efficacy or Comparative Effectiveness by Patient Characteristic or Trauma Type

Each of the five studies examined a different patient characteristic or trauma type and treatment comparison (insufficient SOE for each analysis). The study that compared the efficacy and comparative effectiveness of CPT, PE, and wait-list) between women rape victims with versus without childhood sexual abuse found similar efficacy of CPT and PE across both subgroups (those with versus without childhood sexual abuse).¹²⁵

The second study that compared the efficacy and comparative effectiveness of EMDR, fluoxetine, and placebo between those with childhood-onset (prior to age 18) versus adult-onset trauma⁴⁷ found no significant differences in the efficacy of EMDR as compared with placebo by trauma onset (child vs. adult) as tested by interaction analysis. Interestingly, however, in the main effects analysis, adults with childhood-onset PTSD had worse outcomes than those with adult-onset PTSD, regardless of what intervention they received.

The third study that reported on whether the efficacy of DBT (as compared with a wait-list usual-care comparison group) varied between women with childhood sexual abuse-related PTSD with versus without borderline personality disorder²³ found no differences in efficacy across subgroups. DBT appeared to have similar efficacy, regardless of the presence of comorbid borderline personality disorder.

The fourth and fifth studies examined whether the comparative effectiveness of PE, IPT, and relaxation therapy differed among those with versus without comorbid major depressive disorder,¹³² and trauma type (sexual, physical, or interpersonal), gender, and age of primary trauma exposure (18 years old or younger versus 19 years old or older).¹³⁵ In the first study, the authors reported no significant subgroup differences in PTSD symptom changes at posttreatment in the comparative effectiveness of any of the treatments tested; the effect of the treatment response (CAPS score at posttreatment <20) between groups did not significantly differ at posttreatment. The authors reported that among subjects in the PE treatment group, those with comorbid major depressive disorder had significantly higher attrition rates than those without major depressive disorder (50% vs. 5.6%, respectively, p<0.05).¹³² In the second moderator study, whereas those without sexual trauma exposure had no differences in efficacy across treatment groups, those with sexual trauma exposure had greater efficacy with IPT as compared with PE or relaxation training (p<0.05). Gender and age of primary trauma exposure, however, did not moderate the efficacy findings.

Key Points: Alpha Blockers

• Low SOE supports the efficacy of prazosin on PTSD symptoms reduction.

Key Points: Anticonvulsants

• Low SOE supports the efficacy of topiramate on PTSD symptoms reduction.

Key Points: Atypical Antipsychotics

• Low SOE supports the efficacy of olanzapine and risperidone on PTSD symptoms reduction.
• Low SOE supports the efficacy of risperidone on anxiety symptoms reduction.

Key Points: Selective Serotonin Reuptake Inhibitors

• Moderate SOE supports the efficacy of fluoxetine for PTSD symptoms and low SOE for the anxiety symptoms. Low SOE does not support the efficacy of fluoxetine for depressive symptoms.
• Moderate SOE supports the efficacy of paroxetine for PTSD symptoms, remission, depression symptoms, and disability.
• Low SOE supports the efficacy of sertraline for PTSD symptoms reduction and quality of life (low SOE) and supports no differences in efficacy between sertraline and placebo for depression symptoms.
• Moderate SOE supports no differences in effectiveness between venlafaxine and sertraline for depression and low SOE supports no difference in effectiveness for PTSD symptom reduction, quality of life, and disability.

Key Points: Serotonin and Norepinephrine Reuptake Inhibitors

• Moderate SOE supports the efficacy of venlafaxine for PTSD symptoms, remission, depression symptoms, quality of life, and disability/functional impairment
• Moderate SOE supports no differences in effectiveness between venlafaxine and sertraline for depression and low SOE supports no difference in effectiveness for PTSD symptoms, quality of life, and disability (low SOE for no differences).

Characteristics of Studies

We found three studies meeting our inclusion criteria that studied the efficacy of alpha-blockers (Table 18), each testing the efficacy of prazosin. All enrolled subjects had moderate to severe PTSD. All studies enrolled all or a large majority of male subjects; average age ranged from 30 to 56 years. Trial durations ranged from 8 weeks⁷⁵ to 20 weeks.^{74, 75} Further details describing the included studies are provided in Appendix F.

Table 18

Characteristics of included placebo-controlled trials of alpha-blockers.

Results for Placebo-Controlled Trials of Alpha-Blockers

Characteristics of Studies

Table 19 summarizes the six studies that met inclusion criteria. Appendix F contains further details about each study. The studies enrolled subjects with moderate to severe PTSD. Sample sizes ranged from 28 to 232. Treatment duration ranged from 8 to 12 weeks. Three of the included studies focus on combat-related PTSD;^{77, 164, 165} three enrolled a heterogeneous group of subjects with a variety of index trauma types (e.g., physical and sexual assault/violence, witnessing harm or death, combat, natural disaster, childhood sexual abuse, childhood physical abuse, MVA).^{78, 79, 166} The studies generally recruited middle-aged adults, with mean ages ranging from ~40 to ~55 years. Three studies enrolled at least two-thirds female subjects;^{78, 79, 166} three enrolled all or nearly all males. Five studies used some version of the CAPS as the primary outcome; one assessed PTSD symptoms using the PCL.¹⁶⁵

Table 19

Characteristics of included placebo-controlled trials of anticonvulsants, by drug.

Results of Placebo-Controlled Trials of Anticonvulsants/Mood Stabilizers

PTSD Symptoms

Five of the included studies reported CAPS-assessed PTSD symptom changes between groups (Appendix H). Among the three topiramate studies, only one found significant differences across groups,⁷⁷ although all effect sizes consistently favored topiramate (Figure 19; low SOE).^{78, 79}

Figure 19

Change in CAPS for anticonvulsants compared with placebo. CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; SMD = standardized mean difference; WMD = weighted mean difference.

One study testing divalproex and another testing tiagabine¹⁶⁶ provided insufficient evidence of efficacy for PTSD symptoms due to unknown consistency and imprecise findings.

Characteristics of Studies

Figure 20 summarizes the characteristics of the eight studies meeting our inclusion criteria. Appendix F provides details to further describe the studies included in this section.

Figure 20

Change in CAPS for atypical antipsychotics compared with placebo. CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; PTSD = posttraumatic stress disorder; SIPS = Single Item PTSD Screener; SMD = standardized mean difference.

Evidence of the efficacy of atypical antipsychotics comes from five studies that compared risperidone with placebo and three studies that compared olanzapine with placebo. Relatively small samples (ranging from 15 to 65) tested drug interventions that lasted from 5 weeks to 6 months. Subjects had mean ages generally ranging from 41 to 54. Although two studies enrolled a majority⁸² or only⁸⁴ females, three enrolled exclusively males.^{80, 83, 86, 167}

Table 20

Characteristics of included placebo-controlled trials of atypical antipsychotics, by drug.

Most studies enrolled subjects with trauma types ranging from combat-related trauma,^{80, 83, 85, 86, 167} childhood abuse-related trauma,⁸⁴ mixed types of trauma,⁸² to other types of trauma not related to combat.^{81, 168} One study exclusively enrolled subjects with PTSD and concurrent psychotic features,⁸³ although studies frequently excluded subjects with a history of comorbid schizophrenia, bipolar disorder, or recent substance abuse/dependence.^{81–84, 167}

The majority of studies permitted cointerventions. Most studies assessed PTSD symptoms using some version of the CAPS (CAPS total, CAPS-1, CAPS-2);^80–86 one used the PCL-M to compare between-group outcomes.¹⁶⁷

Results of Placebo-Controlled Trials of Atypical Antipsychotics

PTSD Symptoms

Three studies compared PTSD symptoms between olanzapine and placebo (low SOE). Two studies demonstrated efficacy for CAPS-measured PTSD symptoms (Figure 20).^{80, 81} Another study⁸² that did not use CAPS to assess PTSD symptoms but instead used four other measures of PTSD symptoms, including Single Item PTSD Screeners (SIPS), also favored olanzapine, but differences in this very small study did not reach statistical significance (see Figure 21).

Figure 21

Change in PTSD symptoms for atypical antipsychotics compared with placebo. CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; SMD = standardized mean difference; SIPS = Single Item PTSD Screener.

For risperidone, four studies compared CAPS-assessed PTSD symptoms between treatment and placebo subjects and found some evidence of efficacy (SMD, −0.26, 95% CI, −0.52 to −0.01; I squared=21.1%; 4 studies; N=422; Figure 20; low SOE). One study found no real differences between groups,⁸³ two suggested benefit but found no significant differences in PTSD symptoms between risperidone and placebo,^{84, 85} and one study found modest but very imprecise evidence of risperidone benefit.⁸⁶

Results of Placebo-Controlled Trials of Selective Serotonin Reuptake Inhibitors

PTSD Symptoms

Our meta-analyses found that subjects who received paroxetine, fluoxetine, and sertraline (but not citalopram) had significantly greater decreases in CAPS-assessed PTSD symptoms than subjects who received placebo (Figure 22). The single citalopram study indicated greater decreases in PTSD symptoms among placebo subjects than citalopram subjects, although differences did not reach statistical significance (insufficient SOE). Each of the four fluoxetine studies (five comparisons shown because one study included two fluoxetine arms that compared different doses of the drug) favored fluoxetine (SMD, −0.28; 95% CI, −0.42 to −0.14; I squared=0.0%; moderate SOE). For paroxetine, two studies (one that compared two doses of paroxetine with placebo) each found significant decreases in PTSD symptoms among paroxetine versus placebo subjects (CAPS SMD of −0.56, −0.46, or −0.44 in each study; moderate SOE). Although only three of the seven sertraline studies indicated significant benefit of sertraline for PTSD symptoms, the meta-analysis of pooled data indicated a significant but modest difference of about 5 points on the CAPS between groups (SMD, −0.20; 95% CI, −0.20; 95% CI, −0.36 to −0.04; low SOE). Studies that used other PTSD symptom assessments had consistent findings.^{61, 71, 170, 174}

Figure 22

Standardized mean change from baseline in CAPS for selective serotonin reuptake inhibitors compared with placebo. CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; PTSD = posttraumatic stress disorder; SMD = standardized mean difference; (more...)

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

Eleven of the SSRI studies reported between-group changes in depression symptoms (Figure 23). One small study provided insufficient evidence to determine efficacy of citalopram for reducing comorbid depression in adults with PTSD.¹⁷⁵ The three fluoxetine studies had mixed results with limited evidence of no between-group differences (low SOE for no difference); one study evidenced significant benefit of fluoxetine,⁶¹ another study that tested two different doses of fluoxetine favored both drug arms but not significantly so,⁶² and the third study found the placebo group to have nonsignificantly greater decreases in depression than fluoxetine participants (p=ns).⁴⁷ Both paroxetine studies found significantly greater decreases among intervention group versus placebo group subjects in depression symptoms (moderate SOE).^{64, 65} Decreases in depression symptoms at end-of-treatment did not differ between sertraline and placebo groups (SMD, −0.14; 95% CI, −0.33 to 0.06, 7 studies, N=1,085; low for no difference).

Figure 23

Standardized mean change from baseline in depressive symptoms for selective serotonin reuptake inhibitors compared with placebo. CI = confidence interval; SMD = standardized mean difference; SSRI = selective serotonin reuptake inhibitor.

Four studies assessed the efficacy of SSRIs for anxiety symptoms (Figure 24). Both fluoxetine studies favored the treatment group, but only one significantly so.^{61, 62} The two sertraline studies found effect sizes in the opposite direction, with one study favoring sertraline and the other favoring placebo (insufficient SOE).^{68, 70}

Figure 24

Standardized mean change from baseline in anxiety symptoms (HAM-A) for selective serotonin reuptake inhibitors compared with placebo. CI = confidence interval; HAM-A = Hamilton Anxiety Rating Scale; SMD = standardized mean difference; SSRI = selective (more...)

Characteristics of Studies

Table 22 summarizes the characteristics of the two studies meeting our inclusion criteria. Further details describing the included studies are provided in Appendix F. Both studies evaluated venlafaxine extended release among a heterogeneous group of subjects with a variety of index trauma types. Both studies used CAPS to assess the primary outcome.

Table 22

Characteristics of included placebo-controlled trials of serotonin and norepinephrine reuptake inhibitors.

Results of Serotonin and Norepinephrine Reuptake Inhibitors

PTSD Symptoms

Both included studies reported various measures of PTSD symptoms.^{183, 184} All analyses favored the treatment group, but most comparisons did not indicate significant differences across groups. Overall, we found insufficient evidence to determine the efficacy of either bupropion or mirtazapine for PTSD symptoms (insufficient SOE).

Prevention or Reduction of Comorbid Medical or Psychiatric Conditions

The single bupropion and single mirtazapine studies each favored the treatment group for depression symptoms, but differences did not reach statistical significance across groups (insufficient SOE).^{183, 184} The mirtazapine study reported greater decreases in anxiety symptoms among the treatment versus placebo groups (between-group difference, −1.6, p<0.05; insufficient SOE); the bupropion study did not examine anxiety symptom outcomes.¹⁸⁴

Characteristics of Studies

Table 24 summarizes the four studies that met inclusion criteria. Further details are provided in Appendix F. One study enrolled veterans randomized to paroxetine plus naltrexone (arm not eligible), paroxetine plus placebo, desipramine plus naltrexone (arm not eligible), or desipramine plus placebo.¹⁸⁵

Table 24

Characteristics of included head-to-head pharmacotherapy trials.

Results of Head-to-Head Pharmacotherapy Trials

Network Meta-Analysis of Pharmacotherapy Trials

We conducted network meta-analyses on pre- to posttreatment decreases in PTSD symptoms as measured by the CAPS to determine the comparative effectiveness of drug treatments using all efficacy and comparative effectiveness evidence compiled to answer KQ 2. We did this after checking for any obvious transitivity which would preclude the use of network meta-analysis due to inconsistencies in the sample or treatment characteristics. We did not find any such inconsistencies.

Our network meta-analysis included 33 trials and 13 active treatments (4,491 subjects) that included CAPS-measured PTSD symptoms. A network diagram illustrates the number of subjects contributing to each comparison; thickness of lines connecting each drug-drug or drug-placebo comparison indicates the number of trials with available data for that comparison (Figure 25). We conducted two tests to assess the consistency for the network meta-analysis. First, we compared consistency and inconsistency models that did not differ significantly (χ²(3)=0.06, p=0.997). Next, when possible given the network structure, we tested for differences between direct and indirect estimates using network sidesplits. Direct and indirect estimates did not differ significantly for any of the sidesplit comparisons (Table 25).

Figure 25

Evidence network: comparisons, and number of subjects for each, included in network meta-analysis^a. ^aSample sizes may not add up to those in the diagram because of multiarm studies. CAPS = Clinician-Administered PTSD Scale; N = number; PTSD = posttraumatic (more...)

Table 25

Consistency of network meta-analysis finding.

Figure 26 presents the network meta-analysis findings for each between-group treatment comparison of pre- to posttreatment change in CAPS-assessed PTSD symptoms (SMD and 95% CI displayed for each comparison). We report the findings on the pharmacological interventions for which analyses in the prior section of this report determined at least moderate SOE of efficacy for PTSD symptoms (fluoxetine, paroxetine, and venlafaxine).

Figure 26

Results of network meta-analysis comparing improvement in PTSD symptoms (change in CAPS total score). CAPS = Clinician-Administered PTSD Scale; N = number; CI = confidence interval; PTSD = posttraumatic stress disorder; SMD = standardized mean difference. (more...)

Our network meta-analysis evidenced no significant differences between effectiveness of paroxetine, fluoxetine, and venlafaxine.

Indirect evidence from placebo-controlled trials contributed the majority of evidence to the network meta-analysis. Only four head-to-head comparisons contributed data to the network, none of which compared effectiveness between interventions determined to have at least moderate SOE of efficacy for PTSD symptoms (fluoxetine, paroxetine, and venlafaxine); because indirect comparisons provided the basis for most of the network meta-analysis, we believe the findings only support a low SOE of benefit.

Characteristics of Included Studies

Table 26 summarizes the characteristics of the included study previously described in this report. The study had an end of treatment assessment at 8 weeks and a 6-month followup that included the CAPS as the primary outcome measure of PTSD symptoms. Additional details describing the included study can be found in Appendix F.

Table 26

Characteristics of included pharmacological trials that compared efficacy or comparative effectiveness between subgroups defined by patient characteristics or trauma types.

The analysis conducted to answer this KQ had a small sample size, and therefore did not have the statistical power to detect anything but a very large difference. Many factors other than patient characteristics and trauma type varied across studies, which prevented us from conducting further subgroup analyses using studies that did not, a priori, seek to examine differences in efficacy or comparative effectiveness of interventions across subgroups. Thus, findings should be considered hypothesis generating.

The single study compared EMDR, fluoxetine, and placebo in subjects with a variety of trauma types including child sexual abuse, child physical abuse, child sexual and physical abuse, adult sexual assault, adult physical assault, domestic violence, other adult victimization, traumatic loss, war/terror/violence, and injury/accident.⁴⁷ The authors reported subgroup analyses for those with child-onset trauma and those with adult-onset trauma.

Efficacy or Comparative Effectiveness by Patient Characteristic or Trauma Type

The study compared the efficacy of fluoxetine versus placebo between those with childhood-onset (prior to age 18) versus adult-onset trauma.⁴⁷ No significant index trauma onset by treatment differences was found in the efficacy of fluoxetine compared with placebo by trauma onset (child versus adult) as tested by interaction analysis.

Characteristics of Studies

We found one medium risk of bias study that met our inclusion criteria. Table 27 summarizes the characteristics of the study. Further details are provided in Appendix F.

Table 27

Characteristics of included studies directly comparing psychotherapy with pharmacotherapy.

One study compared subjects with varying trauma types randomized to 8 weeks of fluoxetine, EMDR, or placebo.⁴⁷ Prior KQ 1 and KQ 2 results sections include findings from the placebo comparisons.

Results for Psychotherapy Versus Pharmacotherapy

Characteristics of Included Studies

Table 28 summarizes the characteristics of the one included study previously described in this report. The study had an end of treatment assessment at 8 weeks and a 6-month followup that included the CAPS as the primary outcome measure of PTSD symptoms. Additional details describing the included study can be found in Appendix F.

Table 28

Characteristics of included psychological versus pharmacological trials that examined comparative effectiveness between subgroups defined by patient characteristics or trauma types.

One study compared EMDR, fluoxetine, and placebo in subjects with a variety of trauma types.⁴⁷ The authors reported comparisons between the comparative effectiveness of EMDR versus fluoxetine between those with child- and adult-onset trauma.

Comparative Effectiveness by Patient Characteristics or Trauma Type

The study examined the comparative effectiveness of EMDR versus fluoxetine between those with childhood-onset (prior to age 18) versus adult-onset trauma.⁴⁷ Analyses indicated no significant differences in the comparative effectiveness of EMDR and fluoxetine between those with child- and adult-onset trauma (insufficient SOE).

Characteristics of Studies

The KQ 1 section of the report described brief characteristics of psychological studies rated low or medium risk of bias that included AE assessments. Appendix F contains full information about AE reporting and outcomes for each study.

Characteristics of Studies

The KQ 2 section of the report described brief characteristics of pharmacological studies rated low or medium risk of bias that provided the evidence base for the assessment of AEs of pharmacological interventions.

SSRIs Compared With Placebo