Background: Patient-controlled sedation (PCS) allows the patient to titrate the dosage of sedative drugs according to need. Methods: To compare the efficacy of PCS by using propofol with anesthetist-administered midazolam during colonoscopy, 88 patients were randomized to receive either intravenous midazolam 0.05 mg/kg bolus (1 mg increments as required) or propofol PCS (0.3 mg/kg bolus, zero lockout). Heart and respiratory rates, blood pressure, and oxygen saturation were monitored. Patient cooperation, endoscopist satisfaction, and level of sedation were scored. Patient satisfaction was assessed by questionnaire. To correct for multiple testing of data, statistical significance was asserted only for individually stated p values with p < 0.01. Results: Oxygen saturation and hemodynamics were stable in both groups. Better patient cooperation (good vs. minimal; p = 0.008) and higher endoscopist satisfaction (very good vs. good; p = 0.001) were achieved with PCS. Although more sedated intraoperatively (sedation score 4 vs. 2; p = 0.03 for a single test of hypothesis; correction for multiple testing of data removes this significance), patients in the propofol PCS group were more alert by 30 minutes and discharged earlier (mean and SD times were 43.3 [12.1] min compared with 61.0 [29.7] min; p = 0.001.) More patients in the PCS group (86 vs. 6l%; p < 0.001) were satisfied with their overall level of comfort. Conclusion: PCS with propofol is effective and results in high patient satisfaction and faster discharge. (Gastrointest Endosc 2001;54:8-13.)