There have been a variety of studies in the literature showing that pregabalin is effective for treating generalized anxiety.
The latest of such studies I have seen is published by Mark Boschen in the September 2011 issue of The Canadian Journal of Psychiatry (p.558-565).
This article is a meta-analysis, and shows generally that pregabalin is effective compared to placebo, and has similar, if not greater, effectiveness than other medication options for treating generalized anxiety, such as SSRIs, venlafaxine, and benzodiazepines.
The most common doses have been in the 600 mg/day range, which I consider quite high, particularly since a reasonable dose range for pregabalin could be around 75-300 mg/day.
The "limitations" admitted by the author include issues about dosing, and the fact that Pfizer has funded every published randomized study quoted in the article.
I believe that pregabalin could be a very useful option to try, if a medication trial is being considered for generalized anxiety treatment (of course, the first lines of therapy for generalized anxiety are CBT, relaxation-oriented therapies, meditation, exercise, etc.--but for many people these approaches are not sufficiently helpful). Pregabalin has the advantage of having a quite different--and generally mild--side effect profile compared to other medications, and a what appears to be a fairly low (but I do not think zero) risk of addictiveness/dependence problems, particularly compared to benzodiazepines.
However -- the most obvious limitation of the literature findings is only mentioned briefly in passing by the author in the discussion: it is hard to make a good conclusion about a treatment for anxiety when the duration of follow-up is only 4-8 weeks! I believe that a study for this problem needs to extend for a year or more. First of all, many treatments for anxiety can be acutely helpful, but then wear off substantially over time. Arguably, having a beer every 4 hours could reduce GAD scores over a 4-week trial--but obviously this is not an acceptable long-term treatment! (not only would there be multiple physical harms caused by this over a period of many months or years, but there would be substantial tolerance to anxiety reduction effects, which might only become apparent over many months; furthermore,there would be new psychiatric symptoms induced over a period of months and beyond).
It is not clear from the literature whether the acute benefits over 4-8 weeks from pregabalin would persist over a year or more, whether there would be tolerance, whether there would be longer-term emergent physical or psychiatric side-effects, dependence phenomena, trouble with withdrawal or discontinuation, etc.
Research of this type could be used --spuriously--to justify giving GAD patients benzodiazepines on a routine basis as well, despite the frequent and obvious problem of tolerance, dependence, cognitive problems, etc. Most benzodiazepine studies are of similarly short duration, hence have very limited value to guide us for the long-term treatment decisions which are most important.
Yet, I do think that pregabalin is promising, and could be worth a cautious try, particularly if other approaches are not working well.