Psychomotor abnormalities are one of the complications of opioid therapy in advanced cancer patients. Caffeine has potential properties to counteract the central effects of morphine. Twelve patients receiving stable doses of slow release morphine with adequate pain relief were scheduled for this double-blind placebo-controlled crossover trial. The treatment consisted of an intravenous dose of 1/6 of the daily morphine dose, using an intravenous/oral conversion ratio of 1:3. The dose calculated was administered in 5 minutes. Patients were randomly divided to received in a double-blind manner an infusion of 200 mg of caffeine or saline solution intravenously over one hour. A crossover took place after 2-3 days. Patients were assessed immediately before the infusion and once at the end (one hour after). Each assessment included pain, nausea, confusion, and drowsiness intensity. Psychomotor tests, including tapping speed with 10-30 seconds trials, arithmetic tests, memory for digits, and visual memory were also performed. Caffeine infusion induced a significant decrease in pain intensity (from 25.3 to 16.3, p =0.003), but this was no different from the placebo. Caffeine increased both tapping speed tests (p = 0.041 and 0.010, respectively) in comparison with placebo treatment. No other significant differences were found in the other parameters examined. Caffeine showed a partial effect on the cognitive performance of advanced cancer patients on chronic morphine treatment who received a bolus of intravenous morphine. Further studies are necessary to evaluate whether higher doses of caffeine may be more effective and to establish the role of tolerance to caffeine in this group of patients.