Selective serotonin reuptake inhibitors are the pharmacological treatment of choice for the treatment of social anxiety disorder (SAD). The efficacy and tolerability of fixed doses of escitalopram were compared to those of placebo in the long-term treatment of generalised SAD, using paroxetine as an active reference. Patients with a DSM-IV diagnosis of SAD between 18-65 years of age were randomised to 24 weeks of double-blind treatment with placebo (n = 166), 5 mg escitalopram (n = 167), 10 mg escitalopram (n = 167), 20 mg escitalopram (n = 170), or 20 mg paroxetine (n = 169). Based on the primary efficacy parameter, Liebowitz Social Anxiety Scale (LSAS) total score at Week 12 (LOCF), a significantly superior therapeutic effect compared to placebo was seen for 5 and 20 mg escitalopram and for all doses for the OC analyses. Further improvement in LSAS scores was seen at Week 24 (OC and LOCF), with significant superiority over placebo for all doses of escitalopram, and 20 mg escitalopram was significantly superior to 20 mg paroxetine. Response to treatment (assessed by a Clinical Global Impression-Improvement score < or = 2) was significantly higher for all active treatments than for placebo at Week 12. Clinical relevance was supported by a significant decrease in all the Sheehan disability scores, and the good tolerability of escitalopram treatment. It is concluded that doses of 5-20 mg escitalopram are effective and well tolerated in the short- and long-term treatment of generalised SAD.