Social anxiety disorder (SAD) is a prevalent, disabling disorder. We aimed to assess the effects of pharmacotherapy for SAD and to determine whether particular classes of medication are more effective and/or better tolerated than others. A systematic review and meta-analysis was conducted of all published and unpublished placebo-controlled randomized controlled trials (RCTs) undertaken between 1966 and 2007. A rigorous search, which included searching the Cochrane CCDANTR, MEDLINE and PsycINFO electronic databases, yielded a total of 51 RCTs (9914 participants) considered eligible for inclusion in the review. On average, over half of trial participants responded to medication, as assessed with the improvement item of the Clinical Global Impressions scale (55.2%), with approximately four participants having to be treated for an average of 12 weeks before an additional person responded to medication, relative to placebo (number needed to benefit = 4.19). There was substantial variation across medication classes in the number of dropouts due to adverse events, with an average number needed to harm of 14.4. Maintenance and relapse prevention studies confirm the value of longer-term medication in treatment responders. Medication was also effective in reducing SAD symptoms, comorbid depressive symptoms and associated disability. However, evidence for the efficacy of beta-blockers in treating performance anxiety was lacking. Taken together, trials of selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors provide the largest evidence base for agents that are both effective and well tolerated. This review is an updated version of a Cochrane Review in The Cochrane Library, Issue 4, 2004. Cochrane Reviews are regularly updated as new evidence emerges and in response to feedback, and The Cochrane Library should be consulted for the most recent version of the review.