This pooled analysis of data from two randomized, placebo-controlled trials of venlafaxine extended release (ER) assessed onset of activity and time to response on the 17 symptoms of post-traumatic stress disorder (PTSD) listed in DSM-IV and measured by the 17-item Clinician-Administered PTSD Scale (CAPS-SX17). The intent-to-treat (ITT) population comprised 687 patients (placebo, n=347; venlafaxine ER, n=340). Significant (p<0.05) separation between venlafaxine ER and placebo was observed on most CAPS-SX17 items, with earliest onset of activity and response (week 2) on items 5 (physiological reactivity on exposure to cues) and 14 (irritability or anger outbursts), and (week 4) items 1 (intrusive recollections) and 4 (psychological distress at exposure to cues). Onset of activity and response occurred later (generally, weeks 6–8) on items 9 (diminished interest/participation in activities), 10 (detachment or estrangement), 11 (restricted range of affect), 12 (sense of foreshortened future), all associated with numbing, 15 (difficulty concentrating), 16 (hypervigilance), 17 (exaggerated startle response), associated with hyperarousal, and 6 (avoidance of thoughts/feelings or conversations). Significant differences between venlafaxine ER and placebo were largely absent throughout the treatment period and at the primary week-12 end-point for items 2 (distressing dreams), 7 (avoidance of activities, places or people), 8 (inability to recall important aspect of trauma) and 13 (difficulty falling/staying asleep). These results indicate that symptoms of physiological reactivity and psychological distress in response to cues, and irritability/anger outbursts show early and robust improvement with venlafaxine ER treatment, while symptoms of numbing and hyperarousal take longer. The early and persistent effect of venlafaxine ER over placebo on anger/irritability is noteworthy in view of the clinical significance of these symptoms in PTSD.
Post-traumatic stress disorder (PTSD) is an anxiety disorder characterized by an often heterogeneous set of symptom-behaviours that manifest themselves following exposure to a traumatic event. Symptoms fall into the following three categories (or clusters): re-experiencing (e.g. recurrent and intrusive recollections of the event, flashbacks, distress at reminders of the event); avoidance/numbing (e.g. avoidance of cues associated with the event, inability to recall an important aspect of the trauma, feeling of detachment or estrangement from others); and hyperarousal (e.g. difficulty sleeping, hypervigilance, exaggerated startle response) (APA, 2000). These symptoms must persist for at least 1 month and impair an individual's functioning to meet the criteria for diagnosis (APA, 2000). However, once PTSD is diagnosed, it is often a chronic problem, with a median recovery time of 3–5 yr (Kessler et al., 1995).
The selective serotonin reuptake inhibitors (SSRIs) have become a first-line pharmacological treatment option for PTSD because of their efficacy, tolerability, and safety profiles (Stein et al., 2006; Ursano et al., 2004). Some PTSD symptoms may respond earlier than others to SSRI treatment. In particular, sertraline has been found to result in an early and sustained effect on anger and irritability symptoms in PTSD patients, and to have either no or a weakly significant effect on somatic symptoms such as nightmares, insomnia, physiological distress at exposure to trauma, and exaggerated startle response (Davidson et al., 2002). The early improvement in anger symptoms appeared to mediate the subsequent improvements in intrusive symptoms (Davidson et al., 2002) and was also found to be a significant predictor of the ultimate treatment response with sertraline (Davidson et al., 2004), suggesting that PTSD symptoms that improve early on in treatment may modulate improvement in symptoms that respond later.
The serotonin norepinephrine reuptake inhibitor (SNRI) venlafaxine extended release (ER) has also demonstrated efficacy in the treatment of PTSD (Davidson et al., 2006a,b), as well as mood and anxiety disorders that are commonly comorbid with PTSD, including major depressive disorder (Smith et al., 2002; Thase et al., 2001), generalized anxiety disorder (Allgulander et al., 2001; Davidson et al., 1999; Gelenberg et al., 2000; Rickels et al., 2000), social anxiety disorder (Liebowitz et al., 2005), and panic disorder (Pollack et al., 2007). Understanding the onset and time to response to pharmacotherapy in more detail may provide useful insights into the psychobiology of PTSD as well as clinically useful information. Here we therefore examine the course of PTSD symptom improvement by determining the time to onset and time to response to venlafaxine ER compared with placebo in patients with PTSD, based on pooled data from two studies (Davidson et al., 2006a,b).
A pooled analysis was undertaken combining 12-wk data from two multicentre, flexible-dose, randomized, parallel-group, double-blind studies of venlafaxine ER in the treatment of PTSD. Study 1 was a placebo-controlled, 12-wk study conducted at 59 sites in the USA (Davidson et al., 2006b). Although this study included a sertraline treatment arm, data from the sertraline treatment group are not analysed here. Study 2 was a placebo-controlled, 24-wk study, conducted at 56 sites in 12 countries outside the USA (Davidson et al., 2006a). The study protocol received independent ethics or institutional review board approval before the study began. The study was conducted according to the US Food and Drug Administration Code of Federal Regulations (21CFR, Part 50) and with the ethical principles in the Declaration of Helsinki, and was consistent with Good Clinical Practice and applicable regulatory requirements. Written informed consent was obtained from all subjects before their enrolment. In both studies, patients were randomly assigned to treatment with flexible-dose venlafaxine ER (37.5–300 mg/d) or placebo. Venlafaxine ER dosing was increased to a maximum dose of 75 mg/d at day 5, 150 mg/d at day 14, 225 mg/d at day 28, and 300 mg/d at day 42 (Davidson et al., 2006b).
Participants were male or female outpatients aged ⩾18 yr with a primary diagnosis of PTSD according to DSM-IV criteria with symptoms for ⩾6 months prior to the trial. Futher requirements for participation were a score of ⩾60 on the 17-item Clinician-Administered PTSD Scale, 1-wk Symptom Status Version (CAPS-SX17); a negative serum pregnancy test at screening for women of childbearing potential; generally good health; and a likelihood of complying with the protocol (Davidson et al., 2006a).
The primary trauma types recorded for patients included accidental injury, combat, non-sexual assault, adult sexual assault, childhood sexual abuse, unexpected death, and other (unknown, witnessing, and natural disaster).
Inferential analyses of efficacy variables were performed on the intent-to-treat (ITT) population, which consisted of all randomized patients who received at least one dose of study medication and had at least one post-baseline efficacy assessment (Davidson et al., 2006b). In the individual studies, change in CAPS-SX17 at week 12, with last observation carried forward (LOCF), was measured, along with changes in other outcome measures (Davidson et al., 2006a,b). For this analysis, onset of activity (i.e. significant difference between the venlafaxine ER and placebo groups in mean change from baseline) and time to response (i.e. significant difference between groups in the proportion of patients with item score=0 or 1) on individual CAPS-SX17 items were evaluated using LOCF data for the ITT population at each visit (item score=sum of frequency and intensity scores). Mean changes from baseline CAPS-SX17 item score were compared between the venlafaxine ER and placebo groups using analysis of covariance, with baseline severity as covariate. Logistic regression techniques with baseline severity as covariate were used to compare the proportion of responders (patients with item score=0 or 1). Statistical significance was declared if p⩽0.05.
The ITT population comprised 687 patients (placebo, n=347; venlafaxine ER, n=340). Demographic and baseline clinical characteristics of the study populations are summarized in Table 1. There did not appear to be any clinically important differences in baseline demographic or clinical characteristics between the treatment groups. The majority of patients were female and white. Non-sexual assault was the most frequent trauma type in both the placebo (28.8%) and venlafaxine ER (27.4%) treatment groups. A smaller proportion of patients from the internationally conducted study reported childhood sexual abuse as the primary trauma (1%) compared with those in the United States population (16%). The high mean values for CAPS-SX17 total (>80) are consistent with moderate-to-severe PTSD (Table 2). Scores on the 17-item Hamilton Rating Scale for Depression also suggested moderate depression.
ER, Extended release.
CAPS-SX17, 17-item Clinician-Administered PTSD Scale; CD-RISC, Connor–Davidson Resilience Scale; CGI-S, Clinical Global Impressions–Severity; ER, extended release; HAMD17, 17-item Hamilton Rating Scale for Depression; SDS, Sheehan Disability Scale.
The mean daily dose of venlafaxine ER at the final visit was 203.6 mg; among completers, the mean dose was 225.5 mg. The mean maximum daily dose was 223.1 mg (243.4 mg among completers).
The baseline and week-12 scores (LOCF) for the individual CAPS-SX17 items are presented in Table 3. At week 12, all items, except items 2 (distressing dreams), 7 (avoidance of activities, places or people), 8 (inability to recall important aspect of trauma), and 13 (difficulty falling/staying asleep) improved to a significantly greater degree for patients treated with venlafaxine ER compared with those receiving placebo.
ANCOVA, Analysis of covariance; CAPS-SX17, 17-item Clinician-Administered PTSD Scale; ER, extended release; LOCF, last observation carried forward; PTSD, post-traumatic stress disorder; s.e., standard error of the mean.
p values are based on comparisons of adjusted mean change from baseline scores using ANCOVA.
Table 4 shows time to onset of activity and onset of response for the ITT population (LOCF) on individual CAPS-SX17 items. The baseline prevalence for each item reflects the proportion of patients with scores >2. Venlafaxine ER showed significant (p⩽0.05) separation from placebo on most CAPS-SX17 items on at least one visit. The earliest onset of activity and response (week 2) was observed on items 5 (physiological reactivity on exposure to cues) and 14 (irritability or anger outbursts), and (week 4) items 1 (intrusive recollections) and 4 (psychological distress at exposure to cues).
■, Onset of activity; , onset of response.
Activity, mean CAPS-SX17 item score change from baseline; ANCOVA, analysis of covariance; Baseline prevalence, proportion of patients with CAPS-SX17 item score >2 (on scale of 0–8; item score=sum of frequency and intensity scores), prior to treatment; CAPS-SX17, 17-item Clinician-Administered PTSD Scale; ER, extended release; ITT, intent to treat; LOCF, last observation carried forward; Onset of activity, significant difference between venlafaxine ER vs. placebo groups, analysed by ANCOVA, with baseline severity as covariate; Onset of response, significant difference between venlafaxine ER vs. placebo groups in the proportion of patients with item score=0 or 1, analysed by logistic regression, with baseline severity as covariate; Response, proportion of responders (patients with CAPS-SX17 item score=0 or 1).
p⩽0.05, venlafaxine ER vs. placebo, unless otherwise noted.
Onset of activity and response occurred later (generally, weeks 6–8) on items 9 (diminished interest/participation in activities), 10 (detachment or estrangement), 11 (restricted range of affect), and 12 (sense of foreshortened future), all associated with numbing, 15 (difficulty concentrating), 16 (hypervigilance), 17 (exaggerated startle response), associated with hyperarousal, and 6 (avoidance of thoughts/feelings or conversations).
Statistical separation from placebo on items 2 (distressing dreams), 3 (acting or feeling as if events were recurring), 7 (avoidance of activities, places, people), and 13 (difficulty falling or staying asleep) was somewhat inconsistent. Although venlafaxine ER demonstrated significant difference from placebo for onset of activity with items 2, 3, and 7 and for onset of activity as well as onset of response with item 13, the separation did not persist. Items 2 (distressing dreams), 7 (avoidance of activities, places, people), 8 (inability to recall important aspect of trauma), and 13 (difficulty falling/staying asleep) failed to separate from placebo at week 12. However, item 8 was also the symptom with the lowest baseline prevalence (49.5%) among study participants.
Venlafaxine ER demonstrated a ‘broad spectrum’ effect in the treatment of PTSD, resulting in significant improvements vs. placebo for most of the symptoms examined. The symptoms showing the earliest improvement and response to venlafaxine ER were anger/irritability, physiological reactivity and psychological distress at exposure to cues, and intrusive recollections. Symptoms of numbing (diminished interest/participation in activities, detachment or estrangement, restricted range of affect, and sense of foreshortened future), avoidance of thoughts/feelings or conversations, and hyperarousal (hypervigilance, exaggerated startle response) took longer to respond to venlafaxine ER treatment. Inconsistent efficacy was found for some symptoms, including sleep disturbances (distressing dreams and difficulty falling or staying asleep), acting or feeling as if events were recurring, and avoidance of activities, places, and people. No significant improvement was observed for the inability to recall an important aspect of the trauma; the lack of differential improvement may have been the result of limited room for change for this symptom, since only approximately half of study participants had baseline CAPS-SX17 item scores >2, but could also reflect a more ‘neurobiological’ origin for this symptom. Despite our characterization of the effects of venlafaxine as ‘broad-spectrum’, the distressing nature of these unresolved symptoms, and the need for more effective medications in the future, must be emphasized. Although the study was not powered to find efficacy on particular CAPS-SX17 items, these findings are consistent with the hypothesis that some symptoms respond earlier to pharmacotherapy than others.
This pattern of response to venlafaxine ER treatment resembles that observed with sertraline treatment in PTSD patients, which was associated with an early and robust improvement in anger symptoms that was evident on both onset to activity and time to response measures at every assessment point (Davidson et al., 2002). Similar to the current analysis, the sertraline analysis pooled data from two randomized, placebo-controlled, 12-wk trials that examined the sequence of response of individual PTSD symptoms to treatment. Sertraline was found to exhibit strong effects on anger from week 1 of treatment, which in turn appeared to mediate the subsequent improvements in intrusive symptoms (Davidson et al., 2002). Modulation of anger symptoms has also been found with fluoxetine treatment in combat veterans with PTSD as early as 1 wk after starting treatment (Shay, 1992), and it is possible that the early reduction of anger symptoms by venlafaxine ER seen in this study modulates the later improvements in numbing and hyperarousal symptoms, although we have not examined this possibility. These results, along with the association of serotonin dysregulation with aggression (Bond, 2005; Davis et al., 1997), suggest that antidepressant effects on serotonin play a role in improving anger symptoms in PTSD patients.
In the current studies, venlafaxine ER did not demonstrate a consistent effect on sleep disturbances. Trauma nightmares and sleep disturbances are among the most treatment-resistant symptoms of PTSD, and SSRI treatments have also been found to be generally ineffective for these symptoms (Davidson et al., 2002; Meltzer-Brody et al., 2000). It may therefore be necessary, when using venlafaxine ER or SSRIs to treat PTSD, to supplement these agents with therapies that specifically target sleep disturbances, such as with imagery rehearsal therapy, a hypnotic or hypnotic-like agent (Maher et al., 2006), or an atypical antipsychotic (David et al., 2006; Kozaric-Kovacic et al., 2005; Stein et al., 2002). Prazosin, an α1 adrenergic receptor antagonist, has also recently been found to be an effective treatment for trauma nightmares and sleep disturbances in a randomized, controlled trial of veterans with chronic PTSD (Raskind et al., 2007).
Interpretation of these study results is limited by the relatively short duration of treatment. Although many PTSD patients demonstrate improvements following 12 wk of pharmacotherapy, evidence in patients with major depressive disorder indicates that further improvement occurs with longer treatment times, and non-responders to SSRI treatment at 12 wk have also been shown to become responders following 12–36 wk of therapy (Londborg et al., 2001). In addition, the patients in this study exhibited severe PTSD symptomatology (baseline CAPS-SX17 score >80), and high baseline PTSD severity has been identified as a significant predictor of longer time to treatment response (Londborg et al., 2001). A longer period of observation may therefore be necessary to determine a more complete picture of the sequence of response to venlafaxine ER treatment in this patient population. The study is also limited by sampling characteristics, including the predominance of female patients and heterogeneous trauma types.
Nonetheless, these study results lend further insights into the sequence of response to pharmacological treatment in PTSD, demonstrating that improvements in psychological distress and physiological reactivity after cues, and anger/irritability, occur early on in venlafaxine ER treatment, while numbing and hyperarousal symptoms take longer to respond. It might be speculated that anger/irritability and cue-induced symptoms are mediated by neurocircuitry with serotonergic and noradrenergic receptors, whereas numbing and avoidance involve additional brain regions and neurotransmitters (Charney, 2004). This analytical approach contributes to ongoing attempts to delineate the underlying symptom factors and psychobiological characteristics which contribute to the clinical heterogeneity of PTSD. A forthcoming factor analysis conducted with the same population that was assessed here may help to further characterize how PTSD symptoms cluster in these patients. Future long-term studies of the sequence of symptom response with venlafaxine ER and other treatments in PTSD patients would lend further insights into the recovery process and possible mechanisms through which treatments work in PTSD.
Venlafaxine ER demonstrated a broad spectrum effect in the treatment of PTSD. Symptoms of psychological distress and physiological reactivity in response to trauma cues, and irritability/anger outbursts, showed early and robust improvement with venlafaxine ER treatment, while symptoms of numbing and hyperarousal took longer to respond. The early and persistent effect of venlafaxine ER over placebo on anger/irritability is noteworthy in view of the clinical significance of these symptoms in PTSD. Additional studies examining the psychobiological basis for this sequence of response to pharmacological treatment in PTSD are needed.
This research has been supported by Wyeth Research, Collegeville, PA, USA. The authors acknowledge Xiao Wei Tian, M.S., of Wyeth Research, for performing the statistical analysis on the data described in this paper. The authors also acknowledge Yen-Yee Nydam, Ph.D., Lorraine M. Sweeney, B.A., and Jennifer B. Hutcheson, B.A., of Advogent, a medical communications company that works with Wyeth Research, for their writing and editorial assistance on this manuscript.
Statement of Interest
Dr Stein has received research grants and/or consultancy honoraria from AstraZeneca, Eli Lilly, GlaxoSmithKline, Johnson & Johnson, Lundbeck, Orion, Pfizer, Pharmacia, Roche, Servier, Solvay, Sumitomo, Tikvah, and Wyeth. Dr Rothbaum has received grant support from Janssen and Wyeth. Dr Baldwin has acted as a consultant to Asahi, AstraZeneca, Cephalon, Eli Lilly, GlaxoSmithKline, Lundbeck, Organon, Pfizer, Pharmacia, Pierre Fabre, Roche, Servier, Sumitomo, Wyeth, and has received research grants from Cephalon, Eli Lilly, GlaxoSmithKline, Lundbeck, Organon, Pfizer, Pharmacia, Roche, and Wyeth. Dr Davidson is on the speakers' bureau at Solvay, Pfizer, GlaxoSmithKline, Forest, the Henry Jackson Foundation, the University of Hawaii, the University of Utah, the University of North Carolina, the University of Chicago, the North Carolina Psychiatric Association, the Psychiatric Society of Virginia, Texas University of Psychiatric Physicians, Massachusetts Psychiatric Society, and Duke University Medical School. Dr Davidson has received research support from Pfizer, Eli Lilly, GlaxoSmithKline, Forest, Bristol–Myers Squibb, Cephalon, AstraZeneca, the University of California at Berkeley, Janssen, the International Psychopharmacology Algorithm Project, and the CME Institute. He owns stock in Procter & Gamble, and is an advisor at Actelion, Pfizer, GlaxoSmithKline, Forest, Eli Lilly, Roche, MediciNova, Jazz, AstraZeneca, Wyeth, Sanofi-Aventis, Janssen, Brain Cells, Epix, and Organon. Dr Davidson has also received royalties from MultiHealth Systems, Inc., Guilford Publications, the American Psychiatric Association, Current Medical Science, and Taylor and Francis. Dr Ahmed, Mr Pedersen, and Mr Musgnung are employees of Wyeth Research.