Blanco et al. published this study in the March 2010 issue of Archives of General Psychiatry. Here's a link to the abstract:
Patients with social anxiety were divided into four groups in this randomized prospective 24-week study: placebo; cognitive behavioural group therapy; phenelzine medication; combined CBT + phenelzine.
CBT was modestly effective, phenelzine only slight more effective, but the combination of CBT + medication was substantially more effective, more or less additively so, particularly in terms of total remission rates. There was a very low placebo response.
Findings of this type are not surprising. An interesting aspect to this particular study is that it makes use of phenelzine, an old MAO inhibitor. This shows that sometimes these old drugs can still be quite useful.
This study does not necessarily demonstrate that CBT is the only form of psychotherapy which would work adjunctively to help social anxiety. I do think that components of CBT, such as emphasizing exposure to anxiety-provoking situations, and practicing social initiatives in a systematic way, are necessary. But, other forms of psychotherapy might adjunctively help the CBT to work better!