Opioid analgesics represent one of the most important tools in a sequential pharmacological approach to oncological pain relief. They are recommended by the WHO when nonsteroidal anti-inflammatory drugs (NSAIDs) no longer provide adequate analgesia. However, the use of opioids is limited because of their numerous and often severe adverse effects. This aspect of opioids has motivated continuous research projects aimed at discovering drugs that can provide maximum pain relief but with improved tolerability. Tramadol is a new, centrally acting analgesic with a dual mechanism of action. It shows a selective interaction with mu receptors, which are responsible for nociception, and has weak pharmacodynamic activity on other opioid receptors. At the same time, it acts synergistically on neuroamine transmission by inhibiting synaptic noradrenaline (norepinephrine) reuptake and inducing intrasynaptic serotonin (5-hydroxytryptamine; 5-HT) release. From a pharmacokinetic standpoint, tramadol offers high bioavailability, with similar patterns after oral or parenteral administration (half-life 5 to 7 hours, time to peak plasma concentration 3.1 hours, and approximately 20% plasma protein binding). Although the efficacy of tramadol is comparable to that of other drugs with similar modes of action, the incidence of side effects such as constipation and respiratory depression is lower. The frequency of euphoria and dysphoria is negligible, resulting in little risk of abuse or dependence. It therefore seemed appropriate to further investigate the efficacy and tolerability of tramadol, defined as having only weak potency, in comparison with a widely used opioid, in oncological pain. Buprenorphine was selected as an opioid with a potency equivalent to half that of morphine, but with tolerability that is partially limited by the fact that it frequently gives rise to adverse reactions considered typical of stronger opioids. To compare the analgesic effect and tolerability of tramadol and buprenorphine, 60 patients (44 men, 16 women; average age 61.4 years), all presenting with advanced tumours, were treated orally in a controlled crossover trial with randomised sequences. Patients took both drugs, each for a week, with a 24-hour washout period between treatments. Tramadol was prescribed at the daily dose of 300mg, orally, and buprenorphine at 0.6 mg/day, as a sublingual preparation. Assessments were made of Karnofsky performance status and severity of pain before and during the 4 hours after taking the 2 drugs. Each patient also completed a daily diary recording the severity of pain 1 hour after the dose, the evolution of pain during the day and its severity compared with that on the previous day. They also assessed the duration and quality of sleep. The Karnofsky index changed little with either treatment, but all other variables showed worthwhile improvement, indicating the significant analgesic effect of both drugs. Buprenorphine and tramadol had a similar analgesic effect, although the improvement with the test drug was significant within 1 hour of administration (p < 0.05 compared with baseline) and more marked (p < 0.05 on day 2 compared with buprenorpine). At the end of tramadol treatment, sleep had also improved, both quantitatively and qualitatively (both p < 0.05). The final assessment was significantly in favour of tramadol as regards efficacy (p < 0.05) and patient acceptability (p < 0.01). Thus, tramadol was better tolerated than buprenorphine, and caused fewer and milder adverse reactions. Only 1 patient discontinued tramadol, compared with 18 using reference therapy. Tramadol, although theoretically less potent, nevertheless brought about as much pain relief as the comparator opioid. In conclusion, for this class of drug, tramadol provides an excellent balance between efficacy and tolerability, confirming preliminary studies.