This study was carried out in order to determine if intravenous (i.v.) sedation with diazepam, at the time of procedure, made fibreoptic bronchoscopy more tolerable and if these perceptions persisted on later questioning. Methodology consisted of a sequential, parallel group design comparing sedation with no sedation for bronchoscopy in a tertiary referral hospital. Patient comfort and sedation desired for hypothetical repeat bronchoscopy were assessed both immediately and after at least 1 month. Patients who received sedation rated bronchoscopy as more comfortable (P = 0.01). Those who received sedation were also more likely to want no change in sedation if the bronchoscopy were repeated (P < 0.01). These differences were more evident at later questioning. Sedation was not associated with an increased complication rate but was associated with a prolonged recovery room stay (no sedation, 19.2 min (SEM 3.8) compared with sedation, 76.1 min (5.4), P < 0.001). In contrast to previous studies, our patients found bronchoscopy more comfortable with i.v. diazepam sedation. This was supported by patients who received sedation being less likely to want any change in future sedation if a repeat bronchoscopy were required. The benefit seen with sedation was more marked at later questioning supporting a previously postulated amnesic effect. However, sedation was associated with a prolonged room stay and potentially greater attendant cost.