Eleven patients with cancer pain in a palliative care and chronic pain service required cessation of morphine due to unacceptable opioid side effects. In this retrospective study fentanyl was evaluated as a second-line subcutaneously infused opioid. Starting doses ranged from 100 to 1000 micrograms/24 h, and the duration of fentanyl infusion was 3-70 days. The clinically derived mean relative potency of fentanyl to morphine infusions was 68:1 (SD +/- 23; range: 15-100), and we now recommend cautious dose conversion at an approximate equivalence of 150-200 micrograms fentanyl for 10 mg morphine in non-opioid naive chronic cancer pain patients. All patients demonstrated an improvement in the adverse effect(s) for which the change in opioid was undertaken. Adequate pain relief was achieved in all but 1 patient with mixed nociceptive and neuropathic pelvic pain for whom an epidural infusion of a local anaesthetic/opioid mixture was required. Fentanyl was changed to the more potent synthetic opioid sufentanil in 2 patients for whom the fentanyl dose necessitated too large a volume for the portable syringe driver in use. The clinically derived sufentanil to fentanyl relative potencies were 24:1 and 16:1, respectively. This achieved good analgesia and maintained the favourable side-effect profile seen with fentanyl. Subcutaneous infusion appears to be a safe and viable route of fentanyl delivery, and provided effective analgesia with a low incidence of adverse effects in this small selected group of patients who were intolerant of subcutaneous morphine. We suggest a trial of subcutaneous fentanyl for selected patients who have intractable adverse effects on morphine, and it is now the second-line infusable opioid in our service. Further prospective evaluation of the role of these two synthetic mu opioid agonists in palliative care practice is warranted, as part of an evolving picture of variation in opioid side-effect profile seen with different drugs within the class.