Most current sedative regimens for fibre optic bronchoscopy use an opioid, a benzodiazepine or a combination of both. This study compares midazolam (M) (a benzodiazepine), alfentanil (A) (an opioid) and a combination of both drugs (M+A). One hundred and three patients were randomized in double-blind manner into groups M(35), A(33) and M+A(35). The number of coughs, number of additional aliquots of lignocaine and duration of the procedures were recorded along with oxygen desaturation. The patient's level of discomfort was assessed by patient and bronchoscopist and expressed as a visual analogue score. There were significantly fewer coughs per minute in Group A compared with Group M (P = 0.0053), and significantly less lignocaine was required in Group A (P = 0.005) and in Groups M+A (P < 0.002) compared with Group M. There was no significant difference in the assessment of discomfort between the groups. There was a trend for Group M+A to desaturate more than the other two with a significant difference between desaturation in Group M+A and Group A (P = 0.033). Alfentanil is a more effective anti-tussive agent than midazolam for outpatient fibre optic bronchoscopy. The combination of alfentanil and midazolam does not provide any better anti-tussive effect and may have the risk of a greater degree of desaturation secondary to increased sedation.