Pharmacotherapy in post-traumatic stress disorder: Evidence from randomized controlled trials

Abstract

This review discusses evidence-based pharmacotherapies for post-traumatic stress disorder (PTSD). The epidemiology of PTSD and its phenomenological characteristics are summarized. Focus is placed on the major classes of drugs for which at least a minimum of evidence-based outcome data are available from randomized controlled trials (RCTs). Drugs for the total symptom constellation of the disorder; specific PTSD symptoms, such as nightmares, and prevention of PTSD development post-trauma, are discussed. Where appropriate, RCT methodological problems that limit the conclusions drawn are discussed. In addition, recommendations for research to fill critical gaps in the knowledge of PTSD treatment are offered.

Introduction

Post-traumatic stress disorder (PTSD) is the fourth most common psychiatric disorder in the US, with an estimated lifetime prevalence of 7.8%, which is twice as high in women (10 to 12%) as in men (5 to 6%) [1]. The traumas most commonly associated with PTSD are combat exposure and witnessing injury or death of another among men, and rape and sexual molestation among women. Large epidemiological studies in the US and Australia suggest a high comorbidity for PTSD, with rates of 78 to 80% in women and 85 to 88% in men [1,2]. Data from the National Comorbidity Survey suggests that PTSD is the only anxiety disorder independently associated with both suicidal ideation and suicide attempts, even after controlling for known covariates, such as high levels of comorbidity [3]. Similar to bipolar disorder and major depression, PTSD is among the most impairing of psychiatric disorders, with approximately 63% of patients experiencing severe impairment in at least one of the four role domains measured by the Sheehan Disability Scale [4]. In addition, annual productivity loss from PTSD in the US was estimated to be in excess of US $3 billion in 2001 [5]. PTSD is unique among psychiatric disorders in that there is a clear environmental precipitant, namely the index traumatic stressor requisite to the diagnosis. As detailed by the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV), the types of traumatic exposures (Criterion A1) that may lead to PTSD are varied and include military combat, violent personal assault (physical attack, robbery, mugging and/or sexual assault), being kidnapped or taken hostage, severe vehicular accidents, natural disasters, terror attacks, exposure to or handling dead bodies, and witnessing injury or death of another [6]. In children, sexually traumatic events include developmentally inappropriate sexual experiences, even in the absence of threatened or actual violence or injury. Intentional acts of interpersonal violence, terrorism, sexual assault or combat are more likely to lead to PTSD than accidents, witnessing death or injury, or disasters [1,2,7,8).

Also requisite to the diagnosis of PTSD is the individual's response to the traumatic event, which must involve intense fear, helplessness or horror (Criterion A2) or, alternatively in children, may be characterized by disorganized or agitated behavior [6).

The syndrome of PTSD is defined by three clusters of symptoms that develop after the traumatic exposure and include: re-experiencing the event (Criterion B), avoidance and emotional numbing (Criterion C), and increased arousal symptoms (Criterion D) [6]. These symptoms must persist for > 1 month (Criterion E), and must cause clinically significant distress or functional impairment (Criterion F) [6]. Specification may be made for acute (< 3 months) versus chronic (≥ 3 months) PTSD, as well as for cases with delayed onset of symptoms (≥ 6 months) after the trauma.

Commonly utilized psychometric assessment scales for PTSD are generally based on the DSM-IV definition of PTSD [6]. The Clinician-Administered PTSD Scale (CAPS) has become the gold standard assessment measure for clinical trials, including randomized controlled trials (RCTs) of pharmacotherapies for PTSD [9,10]. Treatment response can also be determined by use of the Clinical Global Impressions scale (CGI) – improvement item (CGI-I) [11] or a similarly defined scale. Self-rated scales utilized in RCTs, particularly for secondary outcomes, have included the Impact of Event Scale (IES) [12] and the Davidson Trauma Scale (DTS) [13].

Literature search, data extraction and analysis

A literature search was conducted in January 2008 using the US National Library of Medicine MEDLINE databases to identify RCTs of pharmacotherapies for the treatment of PTSD. The search was limited to humans, English language articles and publication dates from 1950 to 2008. Keywords and phrases used to search the databases included ‘post-traumatic stress disorder’ and ‘controlled’. The search was supplemented with existing meta-analyses [14-16]. For inclusion, clinical trials were required to: (i) evaluate efficacy for overall PTSD symptomatology, specific PTSD-associated or comorbid symptoms, or prevention of PTSD in recently traumatized (ie, between < 6 h and 14 days following trauma depending on the study criteria) individuals; (ii) utilize randomization; and (iii) utilize a placebo comparator. Of 496 articles returned from the search, 50 articles were accepted for the analysis based on the aforementioned criteria. The analytical goals of this review were to evaluate which agents demonstrated efficacy compared with placebo, provide a critical analysis of methodological characteristics of these trials and provide recommendations for future research.

Pharmacotherapy of PTSD

Current recommendations for the treatment of PTSD include several forms of psychotherapy with demonstrated efficacy [17], as well as psycho-education and supportive measures [18]. Treatment of comorbid conditions, such as mood and substance use disorders, is also an important component of a comprehensive treatment for individuals with PTSD.

Evidence for neurobiological dysfunction in PTSD, such as dysregulation of both main arms of the stress response system, the hypothalamic-pituitary-adrenal axis and the locus coeruleus-norepinephrine system, as well as dysfunctional responses of other neurotransmitter systems, including serotonergic, GABA, and glutamate, suggest a role for pharmacotherapeutic interventions in the treatment of PTSD [18-20]; however, data obtained from RCTs suggest limited efficacy [14-16]. Moreover, the magnitude of positive effects on PTSD symptomatology observed in the RCTs has been small, that is, the relative effect size measurement is < 0.5, prompting the UK's National Institute for Clinical Excellence to recommend that drug treatments should not be used as routine first-line treatments, preferring trauma-focused psychological therapy [15]. In contrast, a meta-analysis conducted by the Cochrane Collaboration of existing RCTs in individuals with PTSD concluded that drug treatments were superior to placebo in reducing the severity of PTSD symptom clusters, as well as comorbid depression and disability [16]. A total of 4,597 individuals from 35 short-term (≤ 14 weeks) RCTs were included in the analysis. There was a superior effect of drug compared with placebo in 13 and 17 of the trials, as measured by the CGI-I response and by a reduction in PTSD symptom severity, respectively. SSRIs represented the drug class evaluated in the greatest number and largest trials to date. A narrative review of three maintenance trials suggested that SSRIs were efficacious in preventing relapse, and longer term treatment may be required for PTSD [16]. These results were consistent with consensus recommendations of 6 to 12 months of pharmacotherapy for acute PTSD and at least 12 months to prevent relapse in chronic PTSD. The British Association of Psychopharmacology has produced guidelines that categorize the evidence for and strength of psychopharmacological recommendations for the treatment of PTSD based on evidence from clinical trials, uncontrolled clinical trials, case reports and expert committee reports [21]. The International Psychopharmacology Algorithm Project proposed algorithms for the psychopharmacological management of PTSD that included treatment sequence when initial choices fail, as well as management strategies in the presence of other comorbidity [22]. In addition, the Veterans Administration and Department of Defense outlined strategies for diagnosis, triage and evidence-based management of both acute stress reactions and PTSD [23].

RCTs of pharmacotherapies have generally targeted the constellation of symptoms that form PTSD, or, more specifically, the symptom clusters of re-experiencing, numbing/avoidance and hyperarousal. Efficacy is generally established when it is demonstrated that the active drug treatment leads to a significantly greater improvement in clinical status compared with placebo, as measured by the CGI-I or similar measure, and total score on a validated PTSD psychometric rating scale, such as the CAPS.

It is also useful to pharmacotherapeutically target one of the particular symptom clusters or individual symptoms observed in PTSD. Comorbid symptoms and conditions of PTSD, such as psychotic symptoms [24], arousal [25], aggression [26], cigarette smoking [27] and alcohol dependence [28] have also been primary targets in some pharmacological trials. Additionally, pharmacological strategies that attenuated memory consolidation after exposure to adversity in preclinical neuroscience studies have been used to prevent the development of PTSD in recently traumatized individuals [29]. Similarly treatments are under development for particular anxiety disorders to enhance the learning and memory processes believed to be involved in the efficacy of particular psychotherapies, such as cognitive behavioral therapy (CBT). Agents that enhance processes of extinction and/or reconsolidation may prove useful in speeding up the response to CBT and prolonged exposure-based, trauma-focused therapies.

Classes of agents investigated for the treatment and prevention of PTSD

Evidence regarding efficacy of the main drug classes evaluated for the treatment of PTSD are summarized in the following sections. In addition to RCTs that assess the efficacy of SSRIs and serotonin-norepinephrine reuptake inhibitors (SNRIs), much of the knowledge gained on PTSD pharmacotherapies is from small, open-label clinical trials. While important for determining which agents should be assessed in larger RCTs, little knowledge can be gained from positive results without a placebo group or comparator. In view of the paucity of RCT data, it is premature to make firm conclusions on the clinical utility of the several diverse classes of pharmacotherapies for PTSD. Moreover, various RCTs for the treatment of PTSD have assessed agents for brief treatment durations, some as little as 4 weeks, thereby limiting the conclusions that can be drawn. Therefore, this review is restricted to specific classes of PTSD pharmacotherapies that have demonstrated efficacy in RCTs. The outcomes of open-label clinical trials, case series and case reports have been disregarded.

The SSRIs

Treatment response to SSRIs has been assessed in a variety of trauma types, including combat, physical and sexual abuse, assault, accident, and witnessing injury or death of another [30-40]. Although the number of participants in clinical trials of SSRIs for the treatment of PTSD have generally been larger than in clinical trials of most other drug classes, several of the SSRI clinical trials have significant methodological problems that substantially limit the conclusions that can be drawn from data obtained. In particular, high overall dropout rates or high differential of dropouts between active drug and placebo, and deficient statistical methodology to handle missing data are common problems.

The methodologically superior and larger RCTs that have demonstrated greater therapeutic effects of SSRI treatment compared with placebo were conducted in mainly (> 90%) civilian populations [30,31,34,37], although not all RCTs were positive [39]. In contrast, two of the larger RCTs that failed to demonstrate any therapeutic benefit of SSRI treatment were in mainly (approximately 71 to 100%) veteran populations with chronic PTSD [32,40], as was the case in a smaller combat veteran (placebo-treated group, n = 6; fluoxetine-treated group, n = 6) clinical trial [33]. This raises the question of whether veterans with chronic PTSD are less responsive to SSRI treatment. In veterans with recent (ie, within several years) exposure to combat, more positive responses have been observed [36,40]. A negative RCT in which the patient cohorts (placebo-treated, n = 10; citalopram-treated, n = 25; sertraline-treated, n = 23) were mainly (≥ 96%) civilians should also be noted [41]. Of the two FDA-approved SSRIs for the treatment of PTSD (sertraline and paroxetine), only paroxetine demonstrated greater efficacy compared with the placebo on all three PTSD symptom clusters [34,37].

A 12-week RCT for the treatment of PTSD in civilian trauma victims (placebo-treated, n = 45; sertraline-treated, n = 49) with comorbid current alcohol dependence assessed alcohol use as the primary outcome measure [28]. Cluster analysis revealed divergent effects based on baseline substance use and PTSD characteristics. Sertraline treatment was associated with fewer drinks per drinking day in the cluster with less severe alcohol dependence and early onset PTSD, whereas placebo was associated with greater decreases in drinks per drinking day and average drinks per day in the cluster with more severe dependence and late onset PTSD [26].

There are several longer-term maintenance clinical trials of SSRIs for the treatment of PTSD [42-45]. In a 24-week, open-label continuation phase after a 12-week RCT, 92% of individuals who responded to sertraline maintained the response in the 6-month continuation phase [44]. A subsequent 28-week, placebo-controlled maintenance phase demonstrated only a 5% relapse in the sertraline group compared with 26% in the placebo group [42]. Moreover, 54% of the acute phase nonresponders became responders in the continuation phase, and high baseline CAPS-2 scores predicted longer time (ie, ≥ 12 weeks) to response. A 6-month RCT of relapse between fluoxetine and placebo demonstrated a relapse rate of 22% for fluoxetine compared with 50% for placebo [43]. In addition, a 24-week continuation phase re-randomized, placebo-controlled relapse prevention trial suggested a lower relapse rate in the fluoxetine group, with 83% of the fluoxetine group and 66% of the placebo group completing the continuation phase [45]. In contrast, a 24-week RCT of paroxetine for relapse prevention was inconclusive as to whether there was a clinically significant effect of paroxetine treatment versus placebo [15]. Therefore, overall, emerging evidence suggests a role for SSRI treatment in PTSD for relapse prevention. These effects suggest that 12-week or shorter clinical trials may be of inadequate duration to observe full treatment response in a subset of patients with PTSD.

SNRIs and other second generation antidepressants

Using the extended release (ER) dual action SNRI venlafaxine, a 6-month RCT was conducted at 56 sites in patients with PTSD (n = 329) [46]. The venlafaxine ER group demonstrated greater reductions in CAPS score, on re-experiencing and avoidance/numbing (but not hyperarousal) cluster scores, and an overall remission rate of 51% compared with 38% for placebo [46]. In a 12-week RCT of venlafaxine ER, sertraline and placebo in patients with PTSD (n = 538), venlafaxine ER improved the 17-item CAPS score compared with placebo [47]. Both active treatments reduced avoidance/numbing, and venlafaxine reduced hyperarousal compared with placebo [47]. Dropout rates in both clinical trials were > 30% and the effect sizes for the main outcome measures were small (~ 0.3). Limitations notwithstanding, the efficacy of venlafaxine ER in the treatment of PTSD was established.

In a 12-week RCT of nefazodone (placebo-treated, n = 15; nefazodone-treated, n = 26), which is both a serotonin reuptake inhibitor and a 5-HT_2A receptor antagonist, the active drug treatment group demonstrated greater changes from baseline in both CAPS total scores and hyperarousal subscale scores compared with placebo [48]. However, first-line use of nefazodone in the treatment of PTSD is unadvisable because of a post-marketing association between this agent and hepatotoxicity [49].

An eight-week RCT of mirtazapine, an α₂-adrenergic antagonist with 5-HT₂- and 5-HT₃-blocking properties, in the treatment of PTSD in a mixed trauma patient group (placebo-treated, n = 9; mirtazapine-treated, n = 17) demonstrated a positive effect on the Short PTSD Rating Interview (SPRINT) global improvement item, but not on total SPRINT scores [50]. Mirtazapine was well tolerated and had positive effects on measures of generalized anxiety.

A small 8-week RCT of sustained-release bupropion was conducted in a mixed trauma patient group (n = 30) with PTSD, some of whom were receiving concurrent SSRIs or neuroleptic treatment [51]. No differences between the group responses were observed on the DTS or CAPS scores, although a group effect was observed for improvement in sleep dysfunction [51].

Tricyclic antidepressants and monoamine oxidase inhibitors

Early clinical trials on the pharmacotherapy of PTSD focused on the tricyclic antidepressants (TCAs) and the irreversible monoamine oxidase inhibitors (MAOIs) [52,53]. The MAOI phenelzine failed to demonstrate benefit in a mixed trauma patient population (10 patients completed 4 weeks of treatment) with PTSD in a small early RCT [54]; however, this clinical trial was conducted for only five weeks. In male veterans with PTSD, both phenelzine and the TCA imipramine were superior to placebo on the IES score at week 5, with greatest improvements observed in the phenelzine treatment group [55]. Combat veterans with PTSD treated with the TCA amitriptyline after 8 weeks demonstrated that active treatment was superior to placebo on the CGI and IES scales [56]. Another early clinical trial in combat veterans with PTSD failed to demonstrate a response to the TCA desipramine on the IES and anxiety ratings [57]; however, this clinical trial was conducted for only four weeks. High dropout rates common in these clinical trials also complicated the interpretation of results.

Two large, multicenter RCTs of the reversible MAOI brofaromine failed to demonstrate significant effects on the symptoms of PTSD [58,59]; however, positive responses on secondary measures in one clinical trial and high placebo response rates in the other trial suggested that future clinical trials may demonstrate therapeutic benefits of brofaromine treatment in PTSD. A combined analysis of the two clinical trials suggested that the predictors of treatment response to brofaromine were lower levels of baseline re-experiencing and avoidance/numbing symptoms, as well as overall less severe PTSD symptoms [60].

Despite promising results from some PTSD RCTs for TCA and MAOI treatment strategies, the characteristics of these classes make them unsuitable to be used as first line agents. For TCAs, there is risk of cardiac rhythm toxicity following overdose, as well as orthostatic hypotension and anticholinergic side effects that limit their use. For MAOIs, the necessity of a low tyramine diet to avoid hypertensive crisis and their contraindication in combination with other antidepressants and sympathomimetics means this class also has limited use. The reversible MAOIs, as well as transdermal delivery of the nonreversible MAOI selegiline, do not require dietary restrictions, providing the possibility of wider use if efficacy was to be established in RCTs.

Anticonvulsants

To date, only a small number of RCTs have investigated the potential use of anticonvulsants for PTSD. A 12-week RCT of lamotrigine in a mixed trauma patient group with PTSD did not demonstrate a significant between-group effect on the Duke Global Rating for PTSD scale [61]. However, the clinical trial was notably underpowered (ten patients completed the trial on active treatment and four on placebo). Improvement was observed for re-experiencing and numbing/avoidance symptom clusters in the lamotrigine group compared with the placebo group.

A small, 12-week RCT of topiramate in patients with non-combat PTSD (placebo-treated, n = 19; topiramate-treated, n = 19) also failed to demonstrate a significant effect on total CAPS scores [62]. In a 7-week augmentation RCT of topiramate in male veterans (randomized patients, n = 40; patients completing the study, n = 24) who were undergoing treatment with other psychotropics, a high dropout rate was reported (55% in the topiramate group compared with 25% in the placebo group), and 40% of patients in the topiramate group experienced adverse events [63]. No treatment effects on primary outcome measures, including CGI and CAPS scores, were observed. A larger, multicenter RCT of tiagabine treatment in a civilian trauma patient group with PTSD (placebo-treated, n = 116; tiagabine-treated, n = 116) failed to demonstrate an effect of active treatment on the CAPS score or secondary outcome measures, despite the sizable patient group [64]. There was also a notably high dropout rate (~ 40%) in the tiagabine treatment group.

An 8-week RCT of divalproex treatment in older (55 ± 7 years) male combat veterans with chronic PTSD (placebo-treated, n = 41; divalproex-treated, n = 41) was conducted, with the CAPS arousal subscale score as the primary outcome measure [25]. No significant difference between divalproex treatment and placebo was observed for reduction of arousal symptoms or other secondary PTSD symptom measures.

To date, there are no RCTs of carbamazepine treatment for PTSD. Therefore, there is little evidence to support the use of anticonvulsants in the treatment of PTSD, although larger RCTs, particularly for lamotrigine, are required.

Benzodiazepines

Benzodiazepines are commonly prescribed for their anti-anxiety and hypnotic properties, but few trials have assessed their effects in PTSD. A 5-week, crossover design RCT of alprazolam in a mixed trauma population with PTSD (patients randomized, n = 16; patients completing the trial, n = 10) failed to demonstrate an effect on PTSD-specific symptoms, although a small positive effect on total Hamilton Rating Scale for Anxiety score was noted [65]. A 1-week RCT of temazepam for recently (ie, 14 ± 10 days) traumatized individuals (placebo-treated, n = 11; temazepam-treated, n = 11) manifesting full criteria for at least two symptom clusters did not demonstrate a group effect on PTSD rates or symptoms at 6 weeks [66].

NMDA receptor modulators

Because of preclinical evidence suggesting a role for the glutamatergic receptor subtype NMDA in fear extinction and reconsolidation, as well as positive effects on cognitive parameters in clinical trials for schizophrenia, d-cycloserine (a partial agonist at a glycine regulatory site of the NMDA receptor) was investigated in a 4-week RCT in civilians with PTSD (randomized patients, n = 11; patients completing the trial, n = 7), most of whom were undergoing treatment with other psychotropics [67]. No significant difference in CAPS score reduction was observed between groups. d-Cycloserine is currently under investigation in RCTs in combination with CBT for the potential to enhance the effectiveness of CBT treatment for PTSD.

Atypical antipsychotics

Several RCTs have investigated the atypical class of antipsychotic agents for the potential treatment of PTSD. In some RCTs, atypical antipsychotics were used as an adjunctive therapy in combination with antidepressant therapy. In other RCTs, the reduction in overall PTSD symptomology was not the main outcome measured. In a 12-week RCT investigating risperidone monotherapy for women with abuse- and assault-related PTSD (randomized patients, n = 20; patients completing treatment, n = 15), the primary visit interaction analysis did not suggest efficacy for PTSD [68]; however, risperidone did lower treatment visit scores on the Treatment Outcomes PTSD Scale compared with placebo. In a 10-week RCT of olanzapine as a monotherapy in a mostly female (14 female and 1 male) abuse- and assault-related PTSD patient group (placebo-treated, n = 5; olanzapine-treated, n = 10), olanzapine had no effect on PTSD [69]; however, a significant weight gain (12 ± 4 lb) was observed in the olanzapine treatment group.

An 8-week RCT of olanzapine augmentation of SSRI treatment was assessed in patients with chronic, military-related PTSD (placebo-treated, n = 9; olanzapine-treated, n = 10) who were minimally responsive to a 12-week SSRI treatment regimen [70]. Olanzapine augmentation treatment produced a significantly greater reduction in CAPS scores; however, the CGI responder status did not differ between groups, which may have been due to the insufficient power of this clinical trial. Improvement in sleep quality ratings correlated with improvements on CAPS score, which suggested much of the improvement in CAPS scores was accounted for by the effect of olanzapine on sleep. The weight gain in the olanzapine group was also substantial (13 ± 6 lb).

Non-remitters (individuals whose total CAPS scores were not reduced by ≥ 70%) in response to 8 weeks of open-label sertraline were investigated in an 8-week RCT (randomized patients, n = 25; patients completing the trial, n = 20) of risperidone augmentation treatment, but no group differences in PTSD symptom reduction were identified [71].

A 4-month RCT of risperidone augmentation (92%) to a stable psychotropic regimen was conducted in male combat veterans with PTSD (randomized patients, n = 65; patients completing the trial, n = 48) [72]. Significantly greater improvements on CAPS total scores and hyperarousal subscale scores were observed for the risperidone treatment group [70]. An 8-week RCT of risperidone augmentation (~ 55%) to a stable antidepressant and/or benzodiazepine treatment regimen or as a monotherapy was conducted in women with chronic PTSD (placebo-treated, n = 9; risperidone-treated, n = 12) related to childhood abuse [73]. The mean CAPS-1 total and subscale scores at 8 weeks did not differ between groups, although significantly greater reductions in CAPS-2 total score, and re-experiencing and hyperarousal subscale scores were observed for the risperidone treatment group, both as an augmentation therapy and as a monotherapy.

A 5-week RCT of risperidone in combat veterans with chronic PTSD and comorbid psychotic features (randomized patients, n = 40; patients completing the study, n = 37) with a primary outcome measure of reduction of psychotic symptoms was conducted [24]. A greater decrease in psychotic symptoms was identified in the risperidone treatment group [24]. The observed reduction in CAPS total scores did not differ between groups. High and differential dropout rates of 42% in the risperidone and 33% in the placebo arm limited the conclusions that could be drawn from these data [24]. A 6-week RCT of low-dose risperidone (0.5 to 2 mg/day) in combat veterans with high hyperarousal on the PTSD Checklist-Military Version (PCL-M) (placebo-treated, n = 8; risperidone-treated, n = 7) demonstrated greater reductions in irritability, and PCL-M cluster B subscale and total scores for risperidone treatment compared with the placebo group [26].

Overall there is little evidence supporting the use of atypical antipsychotics as monotherapies in the treatment of PTSD. In addition, available data on the use of atypical antipsychotics as an augmentation strategy to SSRIs in the treatment of PTSD are equivocal. However, much of the clinical trials have been conducted in treatment-refractory patients with chronic PTSD, suggesting that even limited positive effects are potentially important. It is also notable that concomitant treatments were not controlled in several of these clinical trials, potentially confounding the results. Risperidone may be helpful in selected cases in which comorbid psychotic symptoms are present and, albeit limited, the positive effects on PTSD symptom clusters suggests additional investigation of risperidone is warranted.

Adrenergic inhibitors

Animal models of traumatic stress may have particular relevance in the prevention of the development of PTSD in the aftermath of trauma. Preclinical studies have focused on the role of peripherally administered catecholamines on memory consolidation [74]. Furthermore, noradrenergic signaling in the amygdala is critically involved in integrating the modulatory effects of acute stress on traumatic memories [75]. Inhibition of such noradrenergic modulation has been investigated in an RCT, which aimed to prevent or attenuate the development of PTSD by interfering with memory consolidation in the immediate post-trauma period [29]. The β-adrenergic antagonist propranolol (40 mg/day) was administered to trauma victims (placebo-treated, n = 23; propranolol-treated, n = 18) within 6 h of the trauma [29]. CAPS scores at one-month post trauma were not significantly reduced in the propranolol treatment group compared with the placebo [29]. At three months post trauma, 11% of the propranolol group and 13% of the placebo group had met criteria for chronic PTSD, and a significant preventative effect was not demonstrated [29]; however, the placebo group exhibited greater physiological responses to trauma-script driven imagery [29]. An RCT for PTSD prevention compared propranolol, the anticonvulsant gabapentin and placebo (placebo-treated, n = 17; gabapentin-treated, n = 14; propranolol-treated, n = 17), and treatment was initiated within 48 h of injury at a surgical trauma center [76]. No significant benefit of gabapentin or propanolol treatment was observed compared with placebo for PTSD Checklist-Civilian Version scores, acute stress disorder symptoms at 1 month or PTSD rates at 4 months post-injury [76].

Nightmares and sleep disturbance occur in up to 70% of patients with PTSD, and dysfunction in the noradrenergic system has been suggested to contribute to sleep disturbance [77]. Animal studies have suggested that antagonism at the α₁-adrenergic receptor decreases central noradrenergic neuron firing, which may be of benefit in PTSD-related sleep disturbance. For example, prazosin, an α₁-antagonist, attenuated noradrenergic-mediated suppression of rapid eye movement (REM) sleep [78]. Prazosin has been tested in several RCTs for the treatment of PTSD, in which the impacts on nightmares, insomnia and global clinical status were assessed [77,79,80]. Improvements in all three measures were observed in a 20-week crossover RCT in Vietnam veterans with chronic PTSD (randomized patients, n = 10) [77]. A larger follow-up, 8-week RCT demonstrated prazosin (placebo-treated, n = 17; prazosin-treated, n = 17) improved these measures with large effect sizes (0.94 to 1.08) compared with placebo, and was well tolerated and without significant blood pressure changes [79]. In addition, a small RCT (randomized patients, n = 11) demonstrated that daytime prazosin reduced psychological distress, as measured by the Profile of Mood States total and emotional distress subscale, following challenge with trauma-related words [79]. Daytime prazosin had no significant effects on emotional Stroop completion time [81]. In a patient group with civilian trauma-related PTSD, a small RCT (randomized patients, n = 13) that included home sleep monitoring suggested prazosin-induced reductions in night time PTSD symptoms were accompanied by an increased total sleep time, REM sleep time and mean REM period duration; however, prazosin did not have a sedative-like effect on sleep onset latency [80].

In contrast with prazosin, in an RCT in veterans with chronic PTSD (placebo-treated, n = 34; guanfacine-treated, n = 29), the α₂-adrenergic antagonist guanfacine had no beneficial effects on sleep quality, mood disturbance or PTSD symptoms [82].

Corticosteroids

Neuroendocrine abnormalities identified in PTSD, such as enhanced cortisol suppression to low-dose dexamethasone [83] and the effects of corticosteroids on memory in preclinical models [84], have prompted investigations of peri-trauma corticosteroid administration for the prevention of PTSD. In an RCT (placebo-treated, n = 11; hydrocortisone-treated, n = 9) on the hemodynamic effects of hydrocortisone during septic shock, rates of PTSD diagnosis were assessed at a median timepoint of 31 months after intensive care unit treatment for septic shock [85]. A lower incidence of PTSD was observed in the hydrocortisone treatment group (1/9) compared with the placebo group (7/11). In a larger, single-blinded RCT of hydrocortisone treatment in patients post-cardiac surgery (randomized-treated, n = 91; patients completing the trial, n = 48), treatment was associated with less traumatic memory symptoms at 6 months post surgery [85]; however, the postoperative period of cardiac surgery was not considered a traumatic stressor that would normally fulfill criterion A for PTSD and, therefore, rates of PTSD at 6 months were not compared. Estimated rates of PTSD in response to the stresses of critical illness and critical care are varied [86], making study design of such clinical trials on the use of corticosteroids for PTSD prevention challenging. However, data have been promising and should encourage additional trials defining at-risk populations, determining the risks and benefits of exogenous corticosteroid administration in the critical care setting, and establishing the efficacy of such treatments for prevention of PTSD symptomology.

Conclusions

PTSD is highly prevalent and severely disabling; however, there are few pharmacotherapies with proven efficacy available for PTSD. Current evidence from RCTs supports the use of SSRIs as a first-line treatment option, as SSRIs improve the three core symptom clusters in PTSD in both men and women, and are generally well tolerated. Similarly, the SNRI venlafaxine ER was beneficial for reducing symptoms of PTSD and, therefore, may also be considered a first-line treatment. TCAs and MAOIs may be of benefit in PTSD, but larger and longer duration RCTs are required. Also, safety and tolerability issues with these two classes make first-line use unadvisable. Existing RCT data on anticonvulsant treatment in PTSD suggest these agents are of no benefit, although preliminary data with lamotrigine may warrant conducting a larger RCT. Based on limited RCT data, benzodiazepines have not been effective for core PTSD symptoms. RCTs of atypical antipsychotics have failed to demonstrate their efficacy in the treatment of PTSD as monotherapies and limited evidence supports their use as an SSRI augmentation strategy. The α₁-adrenergic antagonist prazosin was effective for PTSD-related sleep disturbance and nightmares, as well as daytime PTSD-related distress in several small RCTs.

There are several themes in the methodological shortcomings of pharmacotherapy RCTs in the treatment of PTSD. In a number of RCTs, patient group sizes were too small to have sufficient power to detect an effect on overall PTSD symptomology, particularly when the effect sizes determined from the larger trials are often low (~ 0.3). Patient population heterogeneity can also be problematic because of apparent differences in response that depend on chronicity, presence or absence of childhood trauma exposure, comorbidities and trauma types. In general, there has been a failure to separate outcomes based on presenting patient characteristics, such as by trauma type, gender, chronicity of PTSD symptoms and the presence or absence of childhood abuse history. Future clinical trials should strive for greater homogeneity of the patient population for these parameters and/or achieving participant numbers to provide the appropriate statistical power to analyze outcomes based on such subsampies. Potential confounds in pharmacotherapy RCTs, such as in RCTs of the atypical antipsychotics, have often included concomitant drug treatments or psychotherapies. Algorithmic strategies to address partial responses and treatment-resistance are required, but study design of future RCTs should include uniformity across the compared groups with respect to concomitant pharmacological and psychotherapy treatments. Difficulty in detecting effects is compounded when dropout rates are high, making patient retention of great importance, as well as an appropriate a priori statistical plan.

Overall, SSRIs and the SNRI venlafaxine can be recommended as first-line agents to treat the core symptoms of PTSD, while the adrenergic agent prazosin has demonstrated marked efficacy for PTSD-related sleep disturbance and nightmares. Given the great clinical requirement for a more comprehensive pharmacotherapeutic armament for PTSD treatment and prevention, large, extended-duration RCTs for a variety of agents and drug classes are still required.

Table 1.

A summary of evidence-based outcome data from randomized controlled trials.

Medication class	Behavioral target	Medication	Efficacy demonstrated	Number of patients randomized or analyzed (placebo/ treatment)	Reference
SSRIs	PTSD	Paroxetine	Yes	186/365	[34]
				156/151	[37]
				27/25	[35]
		Sertraline	Yes	93/94	[30]
				108/100	[31]
		Fluoxetine	Yes	75/226	[36]
	Remission maintenance	Sertraline	Yes	50/46	[42]
		Fluoxetine	Yes	62/69	[45]
				32/30	[43]
SNRIs	PTSD	Venlafaxine	Yes	168/161	[46]
				179/179	[47]
Alternative second generation ADs	PTSD	Nefazodone	Yes	15/26	[48]
		Mirtazapine	No	9/17	[50]
		Bupropion SR	No	10/18	[51]
MAOIs	PTSD	Phenelzine	Yes	18/19	[55]
		Brofaromine	No	23/22	[59]
				58/56	[58]
TCAs	PTSD	Imipramine	Yes	18/23	[55]
		Amitriptyline	Yes	18/22	[56]
		Desipramine	No	18/18*	[57]
Anticonvulsants	PTSD	Lamotrigine	No	4/10	[61]
		Tiagabine	No	116/116	[64]
		Topiramate	No	19/19	[62]
	CAPS arousal subscale	Valproic acid	No	41/41	[25]
Benzodiazepines	PTSD	Alprazolam	No	10/10*	[65]
	PTSD prevention	Temazepam	No	11/11	[66]
NMDA modulators	PTSD	d-Cycloserine	No	11/11*	[67]
Atypical antipsychotics	PTSD	Olanzapine	No	5/10	[69]
			Yes	9/10	[70]
		Risperidone	No	9/12	[73]
			No	9/11	[68]
	Psychotic symptoms	Risperidone	Yes	18/19	[24]
Adrenergic antagonists	PTSD prevention	Propranolol	No	23/18	[29]
				17/17	[76]
	Insomnia and nightmares	Prazosin	Yes	10/10*	[77]
				17/17	[79]
		Guanfacine	No	34/29	[82]

^*Crossover design with the same patients receiving placebo and active medication. AD Antidepressant, CAPS Clinician-Administered PTSD Scale, MAOI monoamine oxidase inhibitor, PTSD post-traumatic stress disorder, SNRI serotonin and norepinephrine reuptake inhibitor, TCA tricyclic antidepressant.