This study ran from 2004 to 2005 and was approved by the Western Institutional Review Board and subjects provided written informed consent. Subjects were outpatients in the clinic of F.T. who met DSM-IV criteria for PTSD, scored at least 40 on the PTSD Checklist-Civilian Version (PCL-C) (19
), at least 4 (of a maximum of 8) on the Clinician Administered PTSD Scale (CAPS) (20
) “recurrent distressing dreams” item, and at least 4 on the CAPS “difficulty falling asleep/staying asleep” item. Medical history and examination revealed all subjects in good general health and without restless leg syndrome, narcolepsy, and alcohol or other substance abuse for at least three months.
To eliminate subjects likely to develop alpha-1 adrenoreceptor antagonist “first dose hypotension” (21
), each consented subject was prescribed a single blind 1 mg “test dose” of prazosin one week prior to randomization. Of 18 subjects, two experienced subjectively uncomfortable orthostatic dizziness following the “test dose” and were excluded. Three subjects withdrew consent prior to randomization because they decided against research participation (n = 2) or because they left the area (n = 1).
All 13 randomized subjects (11 women; 49 ± 10 years of age [mean age ± SD]) completed ratings in both conditions. Most relevant traumas included childhood sexual abuse (n = 5), childhood physical abuse (n = 3), adult assault (n = 3), rape (n = 1), and life-threatening motor vehicle accident (n = 1). Comorbid psychiatric disorders included major depressive disorder (n = 10, 3 in remission), alcohol abuse in remission (n = 3, mean remission duration = 8.5 ± 7.4 years), obsessive-compulsive disorder (n = 1), and panic disorder (n = 1). Mean PCL-C score at entry was 59 ± 13. All 13 randomized subjects maintained ongoing psychotherapy unaltered through the trial. Eleven subjects were receiving one or more maintenance psychotropic medications. These included sertraline (n = 8, mean dose = 100 ± 41 mg/day), duloxetine (n = 2, mean dose = 70 ± 46 mg/day), and alprazolam (n = 3, mean dose = 0.38 ± 0.18 mg/day). Doses had been stable for at least six weeks and continued unchanged during the trial.
This 7-week 2-period 2-treatment study satisfied criterion for a classic crossover design (22
). Subjects completed random order 3-week treatments of indistinguishable capsules of prazosin and placebo separated by a 1-week washout period. To maintain the blind, random allocation, sequence and medications were determined, distributed and tracked by the study coordinator (PM). Drug was initiated at 1 mg of prazosin (or placebo) at bedtime. Over the next 10 days prazosin (or placebo) were titrated upward by 1 mg increments every two to three days to achieve a therapeutic effect with a minimum of adverse effects. Rapid titration was to enhance subject retention. Prazosin dose achieved (mean 3.1 ± 1.3 mg, range 2–6 mg) or placebo dose achieved (equivalent to 3.2 ± 1.2 mg, range 2–5 mg) were then maintained for 11 days. Phone contact for adverse effect query was made on the morning after drug initiation and after each dose increase through titration phases.
During the last three nights of each treatment, participants wore a REMView (Respironics, Inc.). This battery-operated portable 4.5 ounce device includes head and eye movement sensors that define sleep vs. awake and REM vs. non-REM sleep stages. It has high agreement with polysomnography measures (17
). Total sleep time, REM sleep time and the number of REM sleep periods were generated by REMView software.
Psychiatric and behavioral rating scales (relevant to previous week) were performed by the blinded clinician (FBT) at the first baseline (prior to the first drug trial arm), on the last day of the first drug trial arm, at the second baseline (last day of the washout period) and on the last day of the second drug trial arm. Rating scales included the CAPS “recurrent distressing dreams” (primary measure) and “sleep disturbance” items, a non-nightmare distressed awakenings (NNDA) scale, (a modification of the CAPS “distressing dreams” item that substitutes “NNDA” for “distressing dreams”) (23
), the subject-rated PCL-C, the Clinical Global Impression-Improvement (CGI-I) (24
), and the civilian version of the PTSD Dream Rating Scale (PDRS) (25
) that assesses pathologic PTSD dream content vs. normal dream content. Adverse effects were assessed with the Systematic Assessment for Treatment Emergent Events (SAFTEE) (26
). Sitting heart rate, systolic and diastolic blood pressure were obtained on the last day of each treatment condition.
All three nights of home REMView data in each treatment condition were averaged. Fifty-five nights (29 during prazosin, and 26 during the placebo condition) of interpretable REMView data were available for 10 of the 13 participants. Three subjects’ data were unavailable because of incorrect lead placement, technical failure or battery power loss.
Sample size was estimated by examining the number of subjects used to achieve the effect size in our 2003 crossover study (16
). Repeated-measures analyses of variance (ANOVAs) were performed to evaluate the interaction of change over time with treatment condition. Significance level was p < 0.05. We looked for violations of assumptions of the repeated measures ANOVA test, including order, carryover, and treatment by period effects, and none were found. Effect sizes were calculated using Cohen’s d.