The purpose of this study was to evaluate the safety and advantage of intra-operative patient-controlled sedation (PCS) in elderly patients. Propofol PCS was compared with anesthesiologist-controlled sedation (ACS) during knee arthroplasty under epidural anesthesia. Eleven elderly patients scheduled for unilateral knee total or partial arthroplasty were divided randomly into PCS group (n = 6) and ACS group (n = 5). Epidural anesthesia was performed to produce an appropriate level of sensory block (T 10 through S). Firstly a mixture of pentazocine 0.2 mg.kg-1 and 2% mepivacaine 6-9 ml was injected to the epidural space, and anaesthesia was maintained using 2% mepivacaine afterward. Patients in both groups received propofol 0.3 mg.kg-1 i.v. as a loading dose and 0.6 mg.kg-1.h-1 continuous infusion. Furthermore patients in PCS group received propofol PCS (bolus: 0.2 mg.kg-1, lockout time: 3 min). Patients in ACS group were administered propofol continuously and infusion rates were regulated to maintain a sedation score 3 (Wilson et al) by anesthesiologist. Respiratory rate, blood pressure, heart rate, SpO2, arterial blood gas analysis and plasma levels of propofol were measured 4 times during and after the surgery. Satisfaction of patients and surgeons was questioned. Patients in PCS group received a mean propofol dose of 1.9 +/- 0.1 mg.kg-1 during procedures with a mean duration of 147 min. On the other hand patients in ACS group received propofol 2.9 +/- 0.3 mg.kg-1 with 142 min of procedures. Satisfaction of patients and surgeons, the incidence of complication were similar between the groups. For elderly patients who undergo epidural anesthesia, PCS is a safe and effective technique providing similar good sedation as with ACS.