10 subjects per treatment condition), had a high dropout rate, and did not address missing values (Braun et al., 1990).

Synthesis: The committee found the overall body of evidence regarding benzodiazepines to be scant and low quality. The committee is uncertain about the presence of an effect, and believes that future well-designed studies will have an important impact on confidence in the effect and the size of the effect.

The available evidence is uninformative for the primary use of benzodiazepines in patients with PTSD. The absence of a robust body of evidence regarding benzodiazepines is remarkable in that they are commonly prescribed for patients with PTSD, perhaps as a treatment for anxiety symptoms, while many clinical guidelines recommend against using them at all in this setting (APA, 2004; VA/DOD, 2004). The committee did not examine the evidence regarding the benefits or harms of using benzodiazepines in treating specific symptoms in patients with PTSD.

There are several open-label studies on benzodiazepines, none of which were included, as well as one case-series and one nonrandomized small trial where the patients were treated within approximately 6.7 days after the trauma (range of 2–18 days) so could not have had diagnosed PTSD (Gelpin et al., 1996). There were two trials that focused only on sleep and did not include an overall PTSD outcome that were excluded (Cates et al., 2004; Randall et al., 1995). See Table 3-4 for a summary of the one included clinical trial.

The committee identified four RCTs examining the effects of the monoamine oxidase inhibitors (MAOIs) phenelzine or brofaromine (a selective MAOI not available in the United States) compared with a placebo control (Baker et al., 1995; Katz et al., 1994; Kosten et al., 1991; Shestatzky et al., 1988). Participants in studies had suffered a variety of traumas including combat-related (mostly American former troops), sexual and physical abuse or assault, serious injury, and motor vehicle accidents. The ages in