Fluvoxamine, a selective serotonin reuptake inhibitor, was compared with lorazepam in a multicentre double-blind, parallel group study in 112 general practice patients with mixed anxiety and depression. For inclusion, patients were required to have minimum baseline scores of 21 on the Montgomery-Åsberg Depression Rating Scale (MADRS) and 11 on the Clinical Anxiety Scale (CAS). Treatment was for 6 weeks. There were no significant differences between treatments at any point except in an elderly subgroup in whom anxiety improved more rapidly with lorazepam. There were significant improvements in MADRS, CAS and global ratings compared with baseline at all subsequent assessments. Improvement continued during the whole treatment period. Lorazepam produced more sedation, whilst fluvoxamine produced significantly more nausea and vomiting; this was usually early in onset and, if tolerated, resolved during the course of the study. As it is now widely recognized that benzodiazepines should only be given in short courses of 2–4 weeks, the continued improvement up to 6 weeks has implications regarding choice of treatment.