Posttraumatic Stress Disorder: Acquisition, Recognition, Course, and Treatment

Some of the symptoms of PTSD are considered to be reflections of the adaptive mental processes involved in the assimilation and integration of new information that results from exposure to trauma.⁷ This equates to PTSD symptoms representing part of a normal survival instinct in individuals exposed to trauma, and the pathologic development of PTSD only follows if the response that leads to resolution of the trauma is disrupted in some way. The main conclusion of biological research into PTSD is that exposure to one or more traumatic events triggers a chain of mental and biological events, which ultimately lead to prolonged PTSD.¹⁰

An examination of the longitudinal course of PTSD in a study of rape victims showed that 94% of rape survivors had all the clinical symptoms of PTSD 1 week after the traumatic event, thereby suggesting that such a clinical picture probably constitutes a normal reaction (Figure 1).¹¹ At 3 and 9 months after the event, the proportion of rape victims expressing symptoms of PTSD declines to 15%–25%. Following this period, the curve remains horizontal, indicating that PTSD is an unremitting and chronic disorder that can last for any length of time. In the National Comorbidity Study (NCS), median duration of PTSD associated with worst lifetime trauma was 3 years for those receiving treatment and 5 years for those who did not receive treatment.² However, this estimate does not take into account the real possibility that people may experience PTSD more than once in their lives. Indeed, a great many people report exposure to multiple traumas over a life course.¹² Data from the NCS which used DSM-III-R criteria showed that 60.7% of men and 51.2% of women reported exposure to at least one lifetime traumatic event.² DSM-IV criteria for PTSD expand the set of stressful experiences to include sudden death of a close friend, or loved one, or diagnosis of a life-threatening illness. Results from a community epidemiologic survey in the USA that used DSM-IV criteria reported that 90% of respondents had exposure to at least one lifetime traumatic event.¹³

It has been reported that nonphysically injured trauma survivors admitted to the emergency room, who subsequently went on to develop PTSD, had higher heart rates at the emergency department (mean=95.5 beats per minute, SD=13.9, versus mean=83.3 beats per minute, SD=10.9) (t=4.4, P<0.001) and 1 week later (mean=77.8 beats per minute, SD=11.9, versus mean=72.0 beats per minute, SD=9.5) (t=2.25, P<0.03), but not after 1 and 4 months.¹⁴ Similarly, in a study of trauma survivors, the physiologic response of heart rate, skin conductance and electromyography (frontalis) to mental imagery recorded a short time following the trauma has been shown to differentiate between those who go on to develop PTSD and those who do not (Figure 2).¹⁵ The importance of such physiologic responses is clear, since our clinical experience indicates that PTSD patients can re-access their trauma memories as often as 100 times a day and elicit these physiologic reactions each time. PTSD patients possibly continue to reinforce the initial impact of the trauma by reactivating it in this way. PTSD patients have also been reported to differentiate from normal survivors by poor habituation of skin conductance to a repetition of loud startling noises (Figure 3).¹⁶ Traumatized non-PTSD and anxious groups showed normal habituation by contrast. This may represent a primary defect of the central nervous system that continues to identify and classify the loud tones as threatening in people with PTSD. PTSD patients, therefore, continue to react, rather than rejecting the noises as redundant information and stopping the reaction to them. In a prospective study of 239 trauma survivors,¹⁷ the auditory startle response of all the trauma survivors is normal at 1 week. The response of those patients who go on to develop PTSD becomes abnormal between 1 and 4 months after the trauma, suggesting that this is the critical period during which the central nervous system adapts its response to ambiguous stimuli (such as loud noises) and determines whether PTSD develops.

There are two important questions for the clinician to address when trying to recognize the vulnerable patients that will develop PTSD: 1) Why does trauma lead to PTSD for them rather than some other psychiatric disorder or no disorder at all? 2) What are the risk factors for determining these patients? The acute stress response is universal and nonpredictive of PTSD. Moreover, as mentioned, patients who develop PTSD fail to show a remission of these acute symptoms and show abnormally increased heart rates several days after the trauma as well as other abnormal physiologic responses such as the increased startle response. It would therefore appear that PTSD might develop as a failure of the body to reverse the acute stress response.

The onset of PTSD can be delayed for years. In a large study by Solomon¹⁸ looking at individuals who presented for treatment within 6 years of the Lebanon war, 10% were considered delayed onset, 40% were delayed help-seeking, 33% were exacerbation of subclinical PTSD, 13% were reactivation of recovered PTSD, and the remaining 4% had other psychiatric disorders. This is confirmed in the study by Shalev et al.¹⁹ where 5.1% of patients were truly delayed onset PTSD and the rest were mainly PTSD patients who recovered and were then reactivated by another event.

Most cases of PTSD recover within 1 year, and after 6 years recovery without treatment is unlikely.² However, up to 40% of patients with PTSD have a chronic condition. Chronic PTSD is prolonged and may be unremitting, and subject to reactivation upon exposure to stressors. In addition, it can be disabling and associated with substantial comorbidity. The risk of developing secondary comorbid disorders is related to a number of factors, including the severity of the trauma, gender, family history, past history, and the complexity of the PTSD reaction. Chronic PTSD is linked with abuse of alcohol, drugs, and medication. It is also associated with mortality from suicide.²⁰ The frequency of comorbid disorders with PTSD is shown for male and females in Table 2.² The percentages of individuals with PTSD who have at least one other lifetime disorder is 88.0% for men and 79.0% for women. The major comorbid disorder seen with PTSD is depression, occurring in 47.9% of men and 48.5% of women. Other comorbid disorders include dysthymia, specific phobia and generalized anxiety disorder.

The chronic form of PTSD is often debilitating. The disability associated with PTSD includes work impairment, change in life trajectories, impaired social relations, marital instability and perpetuation of violence. This not only represents a burden to the individual but to society as well.

In a study based on the analysis of the NCS data, which examined the effects of mental disorders on work impairment,²¹ work loss (defined as missing a full day of work), and work cut back (either missing part of a day or working less efficiently than usual) during the previous month was 0.8 days/month and 2.8 days/month, respectively. The amount of work impairment associated with PTSD was the same as that associated with major depression but less than that associated with panic disorder.²¹

In the NCS data, among those with PTSD, there is an increased risk of making suicide plans (odds ratio [OR]=2.4; 95% confidence interval [CI]=1.7–3.3) and an increased risk of attempting suicide (OR=6; 95% CI=3.4–10.7) for patients suffering from PTSD. In addition, marital instability, unemployment, and increased use of outpatient care contribute greatly to the burden to society.¹² The NCS analyses showed that the most extreme adverse effects of traumatic events were associated with complex ongoing traumas that occurred in childhood, such as parental violence, alcoholism, or depression. Such experiences interfere with lifelong patterns of interpersonal relationships and the process of mastering basic educational skills.

In the study by Stein et al.,²² patients with PTSD reported significantly more functional impairment than patients without mental disorders. In addition, patients with PTSD made greater use of healthcare resources than non-mentally-ill patients and encountered considerable functional impairment.

A study of the quality of life with PTSD reported greater impairment at baseline for subjects with PTSD relative to those with major depression and obsessive-compulsive disorder on several domains of the 36-item Short-Form Health Survey.²³ Similarly, in a study of PTSD among civilians, significant impairment was associated with PTSD as seen on the Sheehan Disability Scale, which measures the total work, family and social/leisure disability and the Vulnerability to the Effects of Stress Scale.²⁴

The neurobiology of the acute stress response is summarized in Figure 4. Sensory input from a traumatic event is transmitted to the amygdala, and following cortical input leads to activation of four simultaneous types of response. The first response is the activation of the startle response through the reticularis pontis caudalis. The second response, via the sympathetic nervous system, consists of a release of adrenaline to increase heart rate, blood pressure, blood flow, and increase of glucose to the muscles. This increase in adrenaline is also relevant for the formation of memory. Suppression of the parasympathetic nervous system, which occurs in order to shut down any responses that might compete with the sympathetic nervous system (e.g., digestion, tissue repair, ovulation), constitutes the third response and is independent of the sympathetic nervous system. Finally, in the fourth response, through the hypothalamic-pituitary-adrenal (HPA) axis, the hippocampus and amygdala activate the hypothalamus, which releases corticotrophin release factor (CRF), which in turn activates the pituitary gland to release adrenocorticotrophin hormone. This then stimulates the adrenal gland to release cortisol. The cortisol levels, which are proportional in magnitude to the levels of the stressor, inhibit the sympathetic nervous system, and via a negative feedback inhibition also attenuate the HPA axis.

In healthy individuals, circulating catecholamines and cortisol return to normal within hours. In the long term, exposure to a subsequent stressor (sensitization) can occur and create an exaggerated stress response, but baseline hormone levels will not indicate whether previous stressful events have occurred. However, in individuals with chronic PTSD, both baseline CRF and cortisol levels are decreased. There is also increased sympathetic nervous system activation to trauma, enhanced startle response to both neutral and trauma-related cues, and evidence of decreased parasympathetic nervous system activity. This has led to the view that PTSD may represent an extension of the normal stress response. One hypothesis proposed to explain this is that there are insufficient cortisol levels at the time of the trauma in some individuals, such that the other biologic reactions are not inhibited, and a prolonged activation of some of the stress responses occurs. In support of this, the cortisol levels found in the aftermath of a motor vehicle accident in patients who went on to develop PTSD were found to be lower than in those who remained well or in those who suffered from depression.³⁵ A further study of cortisol levels taken in the emergency room from victims in the aftermath of a rape showed lower levels of cortisol in those who had suffered a previous assault than in those with no previous assault.³⁶ Therefore, prior trauma is definitely a risk for developing PTSD, along with the more common risks such as exposure to trauma, avoidant personality, genetic or familial factors, and cognitive variables such as IQ and education. Studies to isolate which of these variables are particular risk factors for developing PTSD are difficult to design. However, one group of patients that have been shown to be a high-risk group for developing PTSD are the children of Holocaust survivors.³⁷ Further study of this group showed that only the children of parents who developed PTSD during the Holocaust were at risk of PTSD themselves when faced with trauma. Children with PTSD whose parents had PTSD have low cortisol levels, while Holocaust children with no PTSD (and no parental PTSD) have normal cortisol levels. Interestingly, children with no trauma or PTSD but whose parents had PTSD also had low cortisol levels, indicating that the low cortisol may be related to factors that predate the trauma. In short, low cortisol levels may be related to risk for PTSD. The most likely explanation for this low cortisol before trauma is that the cortisol receptors in the pituitary are oversensitive, leading to a greater negative feedback inhibition, which results in lower cortisol levels. This increased receptor sensitivity in PTSD has been successfully demonstrated using a low-dose dexamethasone suppression test, which repeatedly showed an enhanced suppression of cortisol in PTSD patients (Figure 5).^38–42 Therefore, under acute stress, this enhanced negative feedback could result in a premature shut down of the HPA axis, which would prevent the damping down of the sympathetic nervous system. This would lead to a cascade of consequences as detailed in Figure 6. If this cascade model for PTSD is correct, then possible areas for intervention that may help the patient with PTSD become apparent, such as: reduction of nonspecific arousal, reduction of memory-related distress, cognitive restructuring to avoid forming new associations and stimulus generalizations, cognitive restructuring to correct already formed negative associations, addressing any number of the cascade of alterations that have occurred since the trauma (sleep disruption, characterological issues, social avoidance).

Psychosocial treatment of PTSD has shown some promising results and, when effective, has been associated with low relapse rates. Of the existing treatments, exposure therapy has the strongest evidence of efficacy in different populations of trauma victims with PTSD. However, a minority of patients failed to show sufficient gains with this therapy.⁴³ Anxiety management techniques have also been shown to be effective in the treatment of PTSD after rape,¹¹ although not widely studied. Cognitive therapy has also been seen to be effective in rape victims.⁴⁴ More recently, a study by Marks et al.,⁴⁵ of different victim populations with a history of PTSD of at least 6 months' duration studied the treatment effect of prolonged exposure (imaginal and live) alone; cognitive restructuring alone; combined prolonged exposure and cognitive restructuring; or relaxation without prolonged exposure or cognitive restructuring. The study showed that both prolonged exposure and cognitive restructuring were each therapeutic on their own, were not mutually enhancing when combined, and were each superior to relaxation.

Interestingly, this lack of advantage for combining different types of psychosocial treatments was also found in a study by Foa et al.⁴⁶ Studies attempting to examine augmentation of exposure therapy outcome with the addition of cognitive restructuring do not show any additional benefit over exposure therapy alone. No studies have examined the relative efficacy of pharmacotherapy and cognitive-behavioral therapy, and whether combination of the two will augment the efficacy of each.

Eye movement and desensitization reprocessing (EMDR) is a relatively new therapy for PTSD that consists of a form of exposure accompanied by saccadic eye movements.⁴⁷ In EMDR, the therapist asks the patient to visualize images about the trauma, while inducing eye movements by asking the patient to track rapid side-to-side movement of the therapist's finger. A cognitive therapy component is also included by asking patients to replace negative thoughts with positive ones. There have been several studies of the efficacy of EMDR in the treatment of PTSD,^46–49 although most have not been well controlled.⁴³ A well-designed study⁴⁹ found EMDR was inferior to prolonged exposure cognitive therapy.

Successful treatment of PTSD with regard to the patient will involve attitude to treatment, capacity to tolerate distress, availability of support, and extent of comorbidity.⁵⁰ To determine the real benefit of psychosocial treatments for PTSD, further large, well-controlled studies that compare the benefits of specific techniques are required.

The principal goals of pharmacotherapy are reducing PTSD symptoms, improving resilience to stress and quality of life, and reducing disability and comorbidity. Reducing comorbidity is particularly important since patients rarely present with pure PTSD. There are three main classes of drugs that have demonstrated efficacy in the treatment of PTSD: tricyclic antidepressants (TCAs) such as amitriptyline and imipramine, monoamine oxidase inhibitors (MAOIs) such as phenelzine, and SSRIs such as sertraline, paroxetine, and fluoxetine. In addition, a variety of other drugs show promise including antiadrenergics, anticonvulsants and antidepressants (lamotrigine, nefazodone, clonidine).

There is considerable evidence from placebo-controlled trials that TCAs and MAOIs are effective in reducing symptoms of PTSD (Figure 7 and Figure 8).^51–53 There is also preliminary evidence from open-label studies of carbamazepine, valproic acid, and from a small double-blind trial of lamotrigine, that anticonvulsants can produce benefit in PTSD.^54–56

More recently, placebo-controlled studies have shown that the SSRIs sertraline⁵⁷ and paroxetine⁵⁸ are effective in decreasing symptoms of PTSD, and considerable interest has centered on these new data. Similarly, there have been positive placebo-controlled trials for fluoxetine, however results in USA combat veterans^59,60 were not as positive as for the general PTSD population.²⁴ A European study found benefit for fluoxetine in predominantly combat veterans, with relapse prevention effects in maintenance treatment.^61,62

In the 12-week, double-blind study of sertraline in 187 psychiatric outpatients with DSM-III-R PTSD, sertraline treatment yielded a significantly greater improvement than placebo in mean change from baseline for Clinician-Administered PTSD Scale Part 2 (CAPS-2) total score (P=0.02).⁵⁷ In addition, sertraline was significantly better than placebo for the symptom clusters avoidance/numbing (P=0.02) and arousal (P=0.03), but not on reexperiencing/intrusion (P=0.14). In this study, however, 73% of the population was female and 61.5% had suffered physical or sexual assault. Further studies of sertraline are required to demonstrate its efficacy in both genders and across all trauma types. A relapse-prevention study found the risk of relapse to be significantly less with sertraline than with placebo after 9 months of treatment.⁶³

In a 5-week double-blind, randomized, placebo-controlled study of fluoxetine in 64 patients (22 women and 42 men) (31 veteran and 33 nonveterans), fluoxetine significantly reduced overall PTSD symptomatology as assessed by total CAPS score.⁵⁹ However, the difference in PTSD symptom reduction occurred primarily in the subscales “numbing” (for the nonveteran group) and “arousal.” Fluoxetine was an effective antidepressant in the total sample as measured by the Hamilton Depression Rating Scale (HAM-D). However, these improvements in depression did not predict improvement in PTSD score. While there was substantial improvement in depression (HAM-D) in the veteran sample (P= 0.005), there were no meaningful changes in numbing symptoms (P=0.70). Conversely, the nonveteran sample had a modest improvement in depression (HAM-D, P=0.04), but there was a substantial improvement in numbing (P=0.002). In general, nonveteran patients responded much better than veteran patients, which is probably reflective of the higher level of symptomatology in the veterans. In another double-blind, randomized, placebo-controlled study in 53 civilians over a 12-week period, fluoxetine (up to 60 mg/day) significantly reduced overall PTSD symptomatology as assessed by the Duke Global Severity Rating for PTSD, the Structured Interview for PTSD, and the Davidson Trauma Scale (DTS).²⁴ Vulnerability to the effects of stress also responded well to fluoxetine.

In the 12-week, open-label study with paroxetine that was conducted in 19 civilians with DSM-III-R PTSD,⁶⁴ patients showed significantly reduced mean PTSD symptom scores for both the DTS score (P<0.0001) (Figure 9) and the Impact of Event Scale score (P<0.0001). In addition, there were significant improvements in all symptom clusters for both scales. Interestingly, cumulative childhood trauma scores were significantly and negatively correlated with treatment response on the total DTS score (r=0.52, P<0.03). Since the study by Marshall, there has been a series of large clinical trials with paroxetine in PTSD. The clinical program involved 1180 patients with a DSM-IV diagnosis of PTSD, and demonstrated that paroxetine relieves the symptoms of PTSD, significantly reducing mean change from baseline in CAPS-2 total score and Clinical Global Impression global improvement scale. In addition, it significantly reduced the CAPS-2 score for reexperiencing/intrusion (P<0.001), avoidance/numbing (P<0.001), and arousal (P<0.001) compared with placebo. This new data, which was generated in 12-week, double-blind, randomized, fixed- or flexible-dose studies also confirmed that treatment benefit with paroxetine (20–50 mg/day) was observed across all trauma types and in both genders.^65,66

In the fixed-dose study, which was conducted in the United States with 551 PTSD patients who had experienced traumas, including physical or sexual assault, witnessing an accident, experiencing a serious accident/fire/injury, being in war/combat or being involved in a natural disaster, paroxetine was given in doses of either 20 mg or 40 mg/day.^58,66 Decreases from baseline in CAPS-2 total score were significantly greater (P<0.001) for both doses compared with placebo (Figure 10). In addition, decreases from baseline in the 8-Item Treatment Outcome PTSD Scale (TOP-8) total score and DTS total score were also significant for both doses (P<0.001). Similarly, in the U.S. flexible-dose study in which 307 patients received 20–50 mg/day of paroxetine, there were significant decreases from baseline for CAPS-2 total score (P<0.001) (Figure 10), TOP-8 total score (P<0.001), and DTS total score (P<0.001).^58,65

Paroxetine was well tolerated with an adverse event profile similar to that in other anxiety indications.