Forty patients who required day case fibreoptic bronchoscopy were sedated with either nalbuphine 0.2 mg/kg and midazolam 0.05 mg/kg (n = 20), or midazolam 0.05 mg/kg alone (n = 20). Extra midazolam was administered when required. The degree of respiratory depression measured by arterialised venous carbon dioxide levels was recorded together with heart rate, arterial blood pressure, respiratory rate and sedation score, before administration of the drugs and at regular intervals thereafter. Patients who received nalbuphine had slightly higher carbon dioxide levels but respiratory rate and cardiovascular changes were similar in both groups. The addition of nalbuphine to midazolam improves the quality of sedation but prolongs the recovery time and increases the incidence of side effects.