We have assessed the influence of age and preoperative mental status score on postoperative mental function, using a choice reaction time test. One hundred and twelve patients (mean age 70 yr, range 48-88 yr) were given a standardized general anaesthetic for transurethral procedures. Mental status before anaesthesia was measured with the Clifton Assessment Procedure for the Elderly (CAPE). Reaction time was measured before anaesthesia and for up to 3 days after anaesthesia. Variability in reaction time performance was increased 24 h after anaesthesia in transurethral patients, but not in a control group of orthopaedic patients (P = 0.0006). Factors contributing to increased variability of reaction time after operation in a multiple regression analysis were reduced CAPE score before operation (P less than 0.0001), extent of surgery (P = 0.023), postoperative pain (P = 0.007) and postoperative sedative drugs (P = 0.009). Factors not contributing included age, diagnosis of cancer, number of previous operations in past 5 years, duration of anaesthesia, minimum mean arterial pressure, minimum and maximum perioperative PCO2 values, postoperative pyrexia and poor sleep.