The objectives of this study were to identify the characteristics of children who required regional anesthesia for pain associated with terminal malignancy and to identify the safety, tolerability and effectiveness of regional anesthesia as an analgesic modality in terminal pediatric malignancy. A retrospective examination was made of the medical records of children who died of malignancy following treatment at the Dana-Farber Cancer Institute and Children's Hospital, Boston, Massachusetts and who required either epidural or subarachnoid infusions, or neurolytic blockade for pain management (June, 1986--April, 1994) during the terminal phase of their illness. Eleven patients were identified, with a duration of epidural or subarachnoid infusions ranging from 3 days to 7 weeks. Indications for this intervention included limiting side effects of opioids, neuropathic pain unresponsive to either rapid escalation of opioids or massive opioid infusions, analgesia for thoracocenteses for the drainage of malignant pleural effusions and instillation of intrapleural chemotherapy. Pain was localized to one area in ll patients. Analgesia was judged to be satisfactory in all cases after regional anesthesia was instituted and remained satisfactory in all cases throughout the treatment course. Complications associated with regional anesthesia included dural puncture headache and mild respiratory depression. Five patients were nursed at home with either epidural or subarachnoid infusions.