OBJECTIVE To compare, using prilocaine, the effects of continuous spinal anaesthesia (CSA) and continuous epidural anaesthesia (CEA) on haemodynamic stability as well as the quality of anaesthesia and recovery in patients undergoing transurethral resection of the prostate gland. METHODS Thirty patients (>60 yr) were randomized into two groups. Prilocaine, 2% 40 mg, was given to patients in the CSA group, and prilocaine 1% 150mg was given to patients in the CEA group. Incremental doses were given if the level of sensory block was lower than T10 or if needed during surgery. RESULTS There was a significant decrease in mean arterial pressure in Group CEA compared with Group CSA (P < 0.01). The decrease in heart rate in Group CSA occurred 10 min after the first local anaesthetic administration and continued through the operation (P < 0.05). The level of sensory anaesthesia was similar in both groups. The times to reach the level of T10 and the upper level of sensory blockade (Tmax) were 18.0 +/- 4.7 and 25.3 +/- 7.0 min in Groups CSA and CEA, respectively, and were significantly longer in Group CEA. The duration of anaesthesia was 76.8 +/- 4min and was shorter in Group CSA (P < 0.01). CONCLUSIONS Spinal or epidural anaesthesia administered continuously was reliable in elderly patients undergoing transurethral resection of the prostate. Continuous spinal anaesthesia had a more rapid onset of action, produced more effective sensory and motor blockade and had a shorter recovery period. Prilocaine appeared to be a safe local anaesthetic for use with either continuous spinal anaesthesia or continuous epidural anaesthesia.