Seth did an excellent job of parsing out this trial with us during the discussion of this article. I really appreciate a few of his and the group's take-aways, which I will re-iterate here:
1) Consider the medication under study and the authors' big pharma affiliations. Several of the authors (Mellman, Stein, Swift, and Lu) receive fees, stock options, and/or travel support from companies that I am guessing are not manufacturing prazosin, a generic drug. Does this mean the authors receive some secondary gain if they disprove the results of 6 previous RCTs and copious anecdotal and empiric evidence showing that prazosin's effectiveness for PTSD-related nightmares? Who knows. However, we all know that many types of bias creep into biomedical research (see Laura Compton's excellent post on this) and from Freud that many of our drives and motivations lie underneath conscious awareness.
2) After ten weeks clinicians were free to introduce medications and/or psychotherapies that they saw fit to help with their patients' and study subjects' PTSD symptoms with the exception of prazosin or trazodone, or psychostimulants (which should act in the opposite manner to alpha-1 adrenergic antagonists. In my mind this may then have compromised one of the key benefits of the study design, namely the longer trial duration than that of previous RCTs. After 10 weeks this study became more naturalistic in nature. Primary outcomes were scores on the CAPS, PSQI, and CGIC at ten weeks. Scores after ten weeks were considered secondary outcomes and were not listed in the tables. It seems to me as if this trial was really two studies in one, with the junction at ten weeks. We do not know how the introduction of new therapies might have influenced the study results, nor what those therapies were outside of the exclusions.
3) A common critique of this study (mentioned by Seth; Sareen I believe; Kerry Ressler MD, PhD who wrote the editorial accompanying the article; and by Richard Loewenstein MD in the comment at the end of the article) is that the participants were without substance use disorders, active SI, or psychosocial instability. These criteria are not reflective of real-world PTSD patients. Futhermore, SI could be a manifestation of the neurobiological type of PTSD that would respond to prazosin in the first place, so eliminating this symptom may be removing further the very patients whose scores might have responded to prazosin.
4) Finally, prazosin is already widely available in the VA system. If a VA clinician in one of the 12 testing centers felt that their patient would benefit from prazosin, they could just prescribe it for them. They would not enroll that patient in a study in which the patient has a 50% chance of ending up in the placebo group. This is another mechanism by which patients who might have demonstrated response to prazosin never made it into the study.
My future prescribing of prazosin for PTSD-related nightmares will not be affected by the results of this study.
Raskind, M. A. (2018). Pharmacology of Sleep and PTSD: Prazosin - An Alpha-1 Adrenoreceptor Antagonist Approach to Post-traumatic Stress Disorder Pharmacotherapy. Sleep and Combat-Related Post Traumatic Stress Disorder, 349-359. doi:10.1007/978-1-4939-7148-0_30
Ressler, K. J. (2018). Alpha-Adrenergic Receptors in PTSD — Failure or Time for Precision Medicine? New England Journal of Medicine, 378(6), 575-576. doi:10.1056/nejme1716724