Development of a pharmacotherapy for a behavioral disorder is best guided by the pathophysiology underlying its clinical signs and symptoms. Human studies of central nervous system (CNS) noradrenergic activity in PTSD [1–6], laboratory data from an animal model of PTSD [7], and the phenomenology of major PTSD symptoms [8] support increased CNS noradrenergic activity contributing to PTSD pathophysiology. One approach to decreasing excessive CNS noradrenergic activity is to antagonize postsynaptic adrenoreceptors (ARs) in limbic and neocortical areas that are the targets of noradrenergic neurons originating from the locus ceruleus in the rostral pons [9].

Several lines of evidence from studies of sleep physiology, regulation of corticotrophin-releasing factor (CRF) secretion, and noradrenergic modulation of cognition suggest that the postsynaptic alpha-1 AR is a particularly relevant target for pharmacologic alleviation of PTSD symptoms [10–12]. Prazosin is a clinically available alpha-1 AR antagonist that crosses the blood-brain barrier and blocks brain responses to norepinephrine (NE) when administered peripherally [13]. Prazosin has been used safely for decades to treat hypertension and benign prostatic hypertrophy [14, 15]. Randomized clinical trials (RCTs) published over the past 10 years have demonstrated prazosin efficacy for PTSD nightmares, sleep disruption, daytime hyperarousal symptoms, and global clinical status in persons with PTSD from both military and civilian trauma [16–20].

This chapter will review how direct observations of PTSD clinical phenomenology in the context of patient care, together with data from earlier studies addressing CNS noradrenergic activity in PTSD, led to the initial use of prazosin for intractable combat trauma nightmares and sleep disruption in Vietnam combat Veterans; subsequent prazosin RCTs for military and civilian trauma PTSD; prazosin effects on PTSD-like symptoms in a rodent model of PTSD; and the use of prazosin in the treatment of disorders highly comorbid with military PTSD (alcohol use disorder, and persistent postconcussive headaches in military personnel with mild traumatic brain injuries). Finally, it will provide suggestions for optimizing therapeutic effects and minimizing adverse effects of prazosin in the treatment of PTSD and its comorbidities.

Sleep disturbance and recurrent trauma-content nightmares are major hyperarousal and reexperiencing symptoms of PTSD and are prominent among military Veterans [21]. Trauma-content nightmares are a hallmark feature of PTSD [22]. Distressed awakenings and inability to return to sleep with or without recalled trauma nightmares may be the most common symptom motivating combat Veterans to seek treatment for PTSD. In 1996, while providing ongoing clinical care to an African-American Vietnam War combat Veterans group therapy program [23], the author observed that sleep disruption and trauma nightmares were these Veterans’ most troublesome and treatment refractory PTSD symptoms. Psychotropic medications including sedative hypnotics, SSRI antidepressants, and various types of psychotherapies including prolonged exposure [24] rarely had been helpful for these Veterans’ nighttime PTSD symptoms.

Veterans’ accounts of these nighttime PTSD symptoms revealed the following:

CNS-active drugs originally developed to lower blood pressure by reducing noradrenergic activity provide clinically available pharmacologic approaches to reducing excessive CNS noradrenergic activity in PTSD [26]. Using a postsynaptic adrenoreceptor (AR) antagonist rather than a drug that reduced presynaptic NE outflow as a therapeutic approach to intractable and debilitating PTSD trauma nightmares and sleep disruption was inferred from an innovative study performed by Southwick and colleagues in the early 1990s addressing CNS noradrenergic function and its relation to symptom expression in military PTSD [5].

Subject groups were Veterans with PTSD and non-Veteran healthy controls. CNS NE outflow was stimulated by the alpha-2 AR antagonist yohimbine, which eliminates alpha-2 AR autoreceptor inhibition of LC NE release [27]. They measured change in the NE metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) to provide an estimate of CNS presynaptic noradrenergic outflow response, and measured PTSD symptom emergence or worsening to provide an estimate of postsynaptic AR response. Plasma MHPG increased following yohimbine in both PTSD and normal subjects with a somewhat greater increase in PTSD subjects. This modestly greater presynaptic response, however, was overshadowed by the markedly greater postsynaptic PTSD-like behavioral response in the Veterans group. Seventy percent of PTSD Veterans but no control subjects experienced panic attacks, and 40% of PTSD Veterans but no control subjects experienced flashbacks. Total score on a PTSD symptom scale and the individual symptoms emotional numbing, intrusive thoughts, and grief increased substantially in PTSD Veterans but not in controls. These results suggest a major role for increased postsynaptic AR responsiveness in the pathophysiology of PTSD.

There are two major postsynaptic ARs in the human brain, the alpha-1 AR and the beta AR [26]. CNS-active antagonists for both receptors were clinically available. Which postsynaptic AR antagonist should be evaluated for efficacy in PTSD first? The decision to prescribe the CNS-active beta AR antagonist propranolol initially was supported by its previous use in the treatment of other anxiety disorders [28], and case series in which Veterans and children reported PTSD symptom reduction following open-label propranolol treatment [29, 30]. In contrast, no reports of alpha-1 AR antagonist beneficial behavioral effects could be found in the clinical literature. Although the initial choice of a beta AR antagonist proved incorrect (see below), the effects of propranolol on PTSD nightmares paradoxically provided rationale for a trial of prazosin.

The first Vietnam Veteran was treated for intractable PTSD symptoms with a postsynaptic AR antagonist in 1996. He had fought through the bloody TET offensive in 1968 as an Army rifleman with the 1st Infantry Division. He suffered nightly severe trauma nightmares and sleep disruption accompanied by sweating, hypervigilance, and inability to resume sleep. In his graphically realistic nightmares, he repeatedly reexperienced horrific combat traumas. A particularly devastating nightmare “replayed” a terrifying firefight with Viet Cong forces during which a round from the Veteran’s M16 assault rifle accidentally hit and killed his close friend. This recurrent nightmare worsened his intense remorse over this tragic killing, and he developed episodic suicidal ideation and alcohol dependence. Given that his trauma nightmares and sleep disruption had been unresponsive to multiple psychotropic medications and psychotherapies, the author prescribed the postsynaptic beta AR antagonist propranolol 20 mg twice daily (midmorning and before sleep). Two weeks later, the Veteran stated “Doc, we are going the wrong direction. My nightmares are even worse.” That propranolol appeared to increase this combat Veteran’s nightmare intensity was unexpected, but a review of the literature revealed that nightmares are indeed a reported occasional adverse effect of propranolol and other beta AR antagonists [31]. Although disappointing, this unexpected propranolol effect raised the possibility that the wrong postsynaptic AR had been targeted. In several neurobiologic systems, the alpha-1 AR can have opposite effects to those of the beta AR [32, 33]. If the CNS-active beta AR antagonist propranolol worsened this Veteran’s trauma nightmares, would the CNS brain active alpha-1 AR antagonist reduce his trauma nightmares?

The Veteran’s propranolol was discontinued and prazosin initiated at 1 mg at bedtime (drug labeling recommends low-dose initiation and gradual titration of an alpha-1 AR antagonist to avoid “first-dose” orthostatic hypotension) [34]. Prazosin was then titrated upward over 3 weeks to 8 mg at bedtime. At this dose, trauma nightmares ceased, sleep duration increased from 3 to 6 h, and the Veteran reported resumption of “normal” dreaming that had been absent since his Vietnam deployment. This response to prazosin was consistent with two other Veterans’ spontaneous reports of markedly reduced PTSD trauma nightmare intensity and improved sleep following inhibition of prazosin for benign prostatic hypertrophy urinary outflow symptoms. His suicidal ideation also disappeared and he became able to abstain from alcohol. Although the Veteran’s nighttime PTSD symptoms improved substantially, irritability and hypervigilance reemerged every afternoon. Addition of prazosin 5 mg midmorning and midafternoon greatly reduced these daytime hyperarousal symptoms. A modest tachycardia (100–110 beats/min) developed, which was likely a reflex tachycardia response to prazosin. Propranolol was reintroduced at 20 mg bid to treat the tachycardia. It was hoped that the presence of alpha-1 AR antagonism by prazosin would prevent the earlier adverse effect of propranolol to enhance trauma nightmares. Fortunately, trauma nightmares remained absent with combined prazosin and propranolol treatment. It is noteworthy that this Veteran has maintained abstinence from alcohol from 1996 to the present (2016) on a continuous regimen of prazosin and propranolol [35].

The second Vietnam Veteran treated with prazosin had endured 72 days of deadly artillery bombardment and repeated North Vietnamese infantry assaults while besieged with the 26th Marines at Khe Sahn in 1968. His severe and treatment-resistant trauma nightmares, sleep disruption, alcohol dependence, and suicidal ideation were of similar intensity and frequency to those of the first Veteran described above. This second Veteran’s trauma nightmares, sleep disruption, and suicidal ideation also resolved on a regimen of prazosin 5 mg twice daily and 10 mg at bedtime. He too developed a mild reflex sinus tachycardia. Addition of propranolol to prazosin restored normal heart rate without reappearance of trauma nightmares. This second Veteran also has maintained abstinence from alcohol for the past 20 years on continuous maintenance prazosin and propranolol treatment [35].

The apparent beneficial effects of open-label prazosin effects on intractable PTSD symptoms in these and other Veterans under the author’s care led to a search of the literature for studies suggesting alpha-1 AR regulation of neurobiologic systems relevant to PTSD pathophysiology. This search revealed alpha-1 AR modulation of rapid eye movement (REM) sleep, of release of the anxiogenic neuropeptide corticotrophin-releasing factor (CRF), of prefrontal cortex inhibition of primitive cognitive set, and of the acoustic startle response as described below.

PTSD trauma-content nightmares appear to emerge from disrupted REM sleep and from Stage 1 and Stage 2 light sleep [36]. Stimulation of CNS alpha-1 AR disrupts REM sleep, shortens REM sleep duration, and increases Stage 1 and Stage 2 light sleep, effects favoring emergence of trauma nightmares [37–40]. Prazosin reverses these effects [39]. If analogous prazosin effects on sleep physiology occur in humans (see below), they would be consistent with the reported decrease in trauma nightmares and resumption of normal dreaming frequently reported during prazosin treatment of PTSD (unpublished observations).

Studies in rodents and nonhuman primates implicate CRF effects in limbic and neocortical brain areas on the expression of anxiety, fear, and startle [41–44]. Paraventricular nucleus (PVN) CRF neurons are under alpha-1 stimulatory regulation [45, 46]. The increased CRF in cerebrospinal fluid in PTSD is consistent with increased activation of CNS alpha-1 ARs [47, 48]. The alpha-1 AR stimulatory regulation of CRF released from PVN into the portal system to the anterior pituitary is well established; it is likely that alpha-1 AR stimulation also stimulates CRF release from rostrally projecting PVN neurons.

In a series of elegant studies in nonhuman primates, Arnsten and colleagues demonstrated that stimulation of prefrontal cortex high-affinity postsynaptic alpha-2 AR¹ enhances logical thinking and stimulation of lower-affinity alpha-1 AR has the opposite effect [49, 50]. These alpha-2 AR are preferentially activated when NE concentrations are low. In contrast, lower-affinity prefrontal cortex alpha-1 AR are activated under the high NE concentration conditions present during stress. This prefrontal alpha-1 AR stimulation during stress disrupts logical cognitive processing and increases automatic and primitive responses to threats. These cognitive effects of prefrontal alpha-1 AR stimulation are consistent with the cognitive and emotional “fight or flight”-like symptoms of combat trauma PTSD.

The acoustic startle response is under alpha-1 AR stimulatory modulation [51]. Southwick et al. measured acoustic startle following yohimbine or placebo in combat Veterans with PTSD and a comparison group of combat Veterans without PTSD [6]. Only the Veterans with PTSD had an increased acoustic startle response following yohimbine stimulation of noradrenergic outflow. These results from a study with a very appropriate combat-exposed control group suggests that combat trauma PTSD, but not combat exposure per se, is associated with increased alpha-1 AR responsiveness to NE.

The above open-label prazosin treatment observations in combat Veterans together with the alpha-1 AR modulation of neurobiologic systems relevant to PTSD provided rationale for prazosin RCTs for treatment-resistant nighttime PTSD symptoms. Six prazosin randomized controlled trials (RCTs) have been completed and published . Four were performed by our research group at VA Puget Sound and the University of Washington. A fifth was performed at the University of Pittsburgh and VA Pittsburgh. A sixth was performed in Iran by an Iranian and Swiss collaborative group. These studies have demonstrated significant and substantial efficacy of prazosin for reducing nighttime PTSD symptoms reducing daytime hyperarousal symptoms and improving global clinical status.

The participants in the first two RCTs were Vietnam War combat Veterans with decades of treatment-resistant chronic PTSD. Drug was administered as a single evening dose specifically to target persistent and distressing trauma-related nightmares and sleep disruption as primary outcome measures. The Clinical Global Impression of Change (CGIC) [52] also was assessed to determine the impact of nightmare reduction and sleep improvement in global clinical status anchored to function at home and work. Because prazosin duration of action is approximately 6–10 h, the single evening dose regimen in these two studies was not optimally designed to test prazosin effects on daytime PTSD symptoms.

The first RCT was a double-blind placebo-controlled crossover study performed in ten Vietnam combat Veterans [18], all of whom had chronic PTSD with frequent and distressing trauma nightmares. Prazosin or placebo in random order were begun at an initial dose of 1 mg at bedtime and titrated upward for 3 weeks to a dose that eliminated trauma nightmares or to a maximum dose of 10 mg HS. The achieved maintenance dose was maintained for 6 weeks. Following a 1-week washout period, participants were crossed over to the other treatment condition, again for 3 weeks titration and 6 weeks maintenance. At a mean achieved maintenance prazosin dose of 9.6 mg, prazosin was significantly and substantially superior to placebo for reducing nightmares (Clinician-Administered PTSD Scale [CAPS ] “recurrent distressing dreams of the event” item [53]) and sleep disturbance (CAPS “sleep difficulty” item) and improving global clinical status . All Cohen’s d effect sizes for prazosin were large at >1.0. Change in total CAPS score and all three CAPS PTSD symptom clusters (reexperiencing, avoidance, and hyperarousal) also significantly favored prazosin.

The second RCT was a parallel group study . Forty Veterans with chronic PTSD and distressing trauma nightmares were randomized to prazosin or placebo [17]. Most had experienced combat trauma in the Vietnam War. A 4-week dose titration of prazosin or placebo was followed by 8 weeks of maintenance medication (maximum bedtime dose = 15 mg; mean maintenance bedtime prazosin dose = 13.3 mg). Prazosin was significantly and substantially superior to placebo for reducing nightmares and sleep disturbance and improving global clinical status. Effect sizes again were large (Cohen’s d all >0.9). Consistent with the effect of prazosin to increase duration and continuity of REM sleep , dream characteristics of prazosin subjects demonstrated a change from those typical of trauma nightmares toward those typical of normal dreaming as assessed by the PTSD Dream Rating Scale [54]. Although there was a numerically greater reduction in total CAPS score with prazosin than placebo, these differences did not reach statistical significance. Prazosin was well tolerated.

The third RCT was a crossover study in participants with civilian trauma PTSD [20]. This RCT was a collaboration with Fletcher Taylor, MD, a Puget Sound area private practice psychiatrist who had independently observed prazosin beneficial effects on PTSD trauma nightmares in his practice. This study is unique in that it measured effects of both drug and placebo on an objective measure of sleep physiology. Thirteen civilian trauma PTSD participants with severe trauma nightmares and sleep disturbance were randomized to prazosin or placebo in a double-blind crossover trial. Prazosin or placebo was rapidly titrated to 3 mg in the evening during each 3-week treatment period. In the final three nights of each treatment condition, total sleep time, REM sleep time, and sleep latency were recorded at home with the two-lead portable REMView device, which distinguishes sleep from awake state and REM sleep from non-REM sleep [55]. Total sleep time was 94 min longer in the prazosin than in the placebo condition (374 ± 86 vs. 280 ± 105 min, p < 0.01). In contrast, sleep latency (time to fall asleep) was actually several minutes longer in the prazosin condition, consistent with the nonhypnotic nature of prazosin. Both REM time and mean REM period duration were significantly greater during prazosin, suggesting normalization of PTSD disrupted REM sleep. One interpretation of these data is that disruption of REM sleep by inappropriately elevated CNS noradrenergic activity may contribute to the pathogenesis of PTSD trauma nightmares and distressed awakenings. Disrupted REM sleep may also contribute to persistence of excessive emotional response to traumatic memories in PTSD [56]. Improvements in CAPS “recurrent distressing dreams of the event” item scores, PTSD Checklist-Civilian version total scores [57], and global clinical status were significantly greater with prazosin than placebo.

The fourth RCT was performed by Germain and colleagues at the University of Pittsburgh [16]. They randomized 50 Veterans with chronic sleep disturbances to one of three conditions: prazosin (mean dose = 9 mg at night); a behavioral sleep intervention (BSI) that included imagery rehearsal therapy, stimulus control, and sleep restriction; or placebo pill treatment. Both prazosin and BSI were significantly more effective than placebo for sleep improvement, reduction in daytime PTSD symptoms and improvement of global function. Pre- to posttreatment reductions of mean weekly nightmare scores were significantly greater for the BSI and prazosin groups than for the placebo group. Surprisingly, neither active treatment produced significant effects on standard polysomnographic parameters. The efficacy of both prazosin and BSI raises the possibility that a combination of prazosin and BSI, two mechanistically different treatments, may be more effective for PTSD nighttime symptoms than either treatment alone in military Veterans.

The fifth RCT was performed in active duty American soldiers returned from combat deployments in Iraq and Afghanistan [19]. This study is the first prazosin RCT to have prescribed a midmorning prazosin dose in addition to a larger bedtime prazosin dose to increase likelihood of reducing daytime PTSD symptoms. Prescribing this two or three times daily regimen is consistent with prazosin use in general medicine to treat hypertension and benign prostatic hypertrophy. It is also supported by the ability of daytime prazosin to reduce the psychological distress response to trauma cues in PTSD [58]. Sixty seven soldiers in garrison at Joint Base Lewis McChord, Washington, were randomized to prazosin or placebo for 15 weeks. Participants met criteria for PTSD with frequent and severe combat trauma nightmares that had started subsequent to their traumatic combat event(s) in Iraq and Afghanistan. Prazosin was titrated upward over 6 weeks until trauma nightmares were absent or maximum doses of 5 mg midmorning and 20 mg bedtime for men (n = 57) and 2.0 mg midmorning and 10.0 mg bedtime for women (n = 10) were achieved. Maintenance prazosin doses were 4.0 ± 1.2 mg midmorning and 15.6 ± 6.0 mg bedtime for men; and 2.0 ± 0.0 mg midmorning and 7.0 ± 3.5 mg bedtime for women. Prazosin was significantly more effective than placebo for reducing CAPS “recurrent distressing dreams of the event” item scores; Pittsburgh Sleep Quality Index [59] scores; and total 17-item CAPS scores (reduction from baseline = 25.1 ± 3.4 prazosin group and 13.8 ± 3.3 placebo group [(p = 0.02]). Total CAPS score decrease remained significantly greater in the prazosin group (p = 0.04) even after removing the nightmare item. The proportion of treatment “responders,” defined as CGIC ratings “moderately improved” or “markedly improved” in ability to function at home and at work, was 64% for the prazosin group and 27% for the placebo group (p < 0.001). Even at the relatively high doses achieved, prazosin was well tolerated by these young soldiers. This study demonstrated that prazosin has clinically meaningful beneficial effects on both daytime and nighttime PTSD symptoms in active duty combat experienced soldiers when administered twice daily. Similar open-label prazosin beneficial effects with good tolerability have been reported in soldiers performing combat operations in the dehydrating Iraq desert warfare environment [60] and in elderly World War II Veterans and Holocaust survivors [61].

A sixth RCT was performed in 100 Iranian PTSD patients (28% women) equally divided between military Veterans and victims of civilian trauma [62]. Interestingly, subjects were randomized to one of three conditions: prazosin titrated to 15 mg at bedtime, the sedating antihistamine hydroxyzine titrated to 100 mg at bedtime, or placebo. Although sleep quality was the primary measure of interest, overall PTSD symptoms were assessed as well.

Improvement in Pittsburg Sleep Quality Index scores was significantly greater in both prazosin subjects (15.5 ± 2.2 to 10.2 ± 2.1) and hydroxyzine subjects (15.6 ± 2.2 to 12.2 ± 2.1) than placebo subjects (15.5 ± 2.0 to 15.0 ± 1.7). Both prazosin and hydroxyzine also were significantly superior to placebo for overall PTSD score and nightmare score, with prazosin significantly superior to hydroxyzine for overall PTSD and nightmares.

Finally, a small (n = 8) 16 day per treatment arm crossover study of the alpha-1 AR antagonist doxazosin for PTSD in military Veterans was recently reported [63]. Doxazosin has a substantially longer duration of action than prazosin, and an animal study suggests it penetrates into brain [64]. Doxazosin was not superior to placebo on the CAPS but did demonstrate a significant albeit modest superiority to placebo on the PTSD checklist.

In the above described RCT of prazosin for PTSD in active duty soldiers [19], neither pretreatment PTSD symptom pattern nor severity distinguished between the 64% of soldiers who globally benefited from prazosin and the 36% who did not. The most direct potential biomarker for predicting prazosin efficacy would be the presence of elevated brain alpha-1 AR activity, the presumed therapeutic target of prazosin action. Although brain alpha-1 AR activity unfortunately is not currently measurable, brain and peripheral alpha-1 activity are co-regulated [65]. Standing systolic blood pressure (BP) is an easily measured and reliable function of alpha-1 AR activity on peripheral arterioles [66]. In a secondary analysis using linear mixed-effects models, we demonstrated a strong effect of standing systolic BP on PTSD outcome measures [67] in soldiers randomized to prazosin. Each 10 mmHg higher baseline systolic BP increment resulted in an additional 14-point reduction of CAPS total PTSD symptom score at endpoint (p = 0.002). For example, a higher standing systolic BP of 130 mmHg predicted a large 34-point CAPS score reduction, whereas a lower standing systolic BP of 110 mmHg predicted only a 7-point CAPS score reduction (numerically smaller than the mean placebo group response). All other combinations of BP parameters and PTSD outcome measures were similarly significant (although less robust than standing systolic BP) or demonstrated trends in the predicted direction. In the placebo group, there was no signal for a baseline BP effect on PTSD outcome measures. These findings suggest that a relatively high standing systolic BP for a person’s demographic group is a potential biomarker that helps identify persons with PTSD more likely to respond to prazosin. They also are consistent with high alpha-1 AR activation contributing to the pathophysiology of PTSD in substantial subgroup patients.

We assessed the effects of pharmacologic reduction of CNS noradrenergic activity on PTSD-like behaviors in a rodent model of PTSD [7]. Mice were exposed to foot shock followed by three weekly contextual reminders of the shock exposure. This modification of a rodent model of PTSD developed by Pynoos [68] produced a group of “susceptible” mice that developed increased acoustic startle response and aggression and decreased social interaction and a group of “resilient” mice that did not differ on these behavioral parameters from control mice. Both antagonism of NE effects at postsynaptic alpha-1 AR with prazosin and reduction of presynaptic NE outflow with the alpha-2 inhibitory autoreceptor agonist clonidine normalized acoustic startle response, aggression, social interaction, and other PTSD-like behaviors in susceptible mice. These data are consistent with both increased postsynaptic alpha-1 AR responsiveness and increased presynaptic NE outflow contributing to PTSD pathophysiology.

Although effects of a beta AR antagonist on established PTSD-like behaviors in an animal model (or in humans) have not been examined empirically, there is theoretical support for beta AR antagonism as a potential prophylactic intervention to reduce PTSD incidence when administered immediately following trauma [69]. Unfortunately, clinical studies addressing this hypothesis have overall been equivocal or negative [70, 71]. A recent PTSD propranolol prophylaxis study in rats is consistent with these disappointing clinical studies [72]. This study assessed the ability of a single post-trauma dose of propranolol to prevent expression of PTSD-like behaviors. An hour after exposure to well-soiled cat litter, rats were administered single bolus subcutaneous propranolol or normal saline. Propranolol effectively slowed heart rate and impaired memory on the object recognition task, indicating successful antagonism of both peripheral and CNS beta AR. However, 30 days following trauma exposure, there were no propranolol effects on PTSD-like behaviors.