BACKGROUND: Colonoscopy is generally performed with the patient sedated and receiving analgesics. However, the benefit of the most often used combination of intravenous midazolam and pethidine on patient tolerance and pain and its cardiorespiratory risk have not been fully defined. METHODS: In this double-blind prospective study, 150 outpatients undergoing routine colonoscopy were randomly assigned to receive either (1) low-dose midazolam (35 micrograms/kg) and pethidine (700 micrograms/kg in 48 patients, 500 micrograms/kg in 102 patients), (2) midazolam and placebo pethidine, or (3) pethidine and placebo midazolam. RESULTS: Tolerance (visual analog scale, 0 to 100 points: 0 = excellent; 100 = unbearable) did not improve significantly more in group 1 compared with group 2 (7 points; 95% confidence interval [-2-17]) and group 3 (2 points; 95% confidence interval [-7-12]). Similarly, pain was not significantly improved in group 1 as compared with the other groups. Male gender (p < 0.001) and shorter duration of the procedure (p = 0.004), but not amnesia, were associated with better patient tolerance and less pain. Patient satisfaction was similar in all groups. Oxygen desaturation and hypotension occurred in 33% and 11%, respectively, with a similar frequency in all three groups. CONCLUSIONS: In this study, the combination of low-dose midazolam and pethidine does not improve patient tolerance and lessen pain during colonoscopy as compared with either drug given alone. When applying low-dose midazolam, oxygen desaturation and hypotension do not occur more often after combined use of both drugs. For the individual patient, sedation and analgesia should be based on the endoscopist's clinical judgement.