In the past, pain control for chronic pain syndromes using narcotic infusion has been carried out primarily via the intrathecal (subarachnoid) route. This report presents one of the first large series of terminally ill cancer patients with intractable pain treated with continuous epidural morphine infusions by means of implanted pumps and epidural spinal catheters. The purpose of the study was to demonstrate that the epidural route is effective with minimal complications, and that screening with temporary epidural catheter infusions results in a high rate of subsequent pain relief. A multidisciplinary team (neurosurgeon, anesthesiologists, psychiatrists, oncologists, and nurse clinicians) evaluated and treated all of the patients studied. Percutaneous placement of temporary epidural catheters for a trial assessment was performed by the anesthesiologists. Pain evaluations were conducted independently by psychiatrists using both verbal and visual analog scales. From 1982 to 1988, 41 (59.4%) of 69 patients evaluated for eligibility experienced good pain control during trial assessment and were subsequently implanted with Infusaid infusion pumps. Preinfusion pain analog values were 8.6 +/- 0.3 and postimplantation values at 1 month were 3.8 +/- 0.4 (p less than 0.001). Over this same 1-month period. requirements of systemic morphine equivalents decreased by 79.3% with epidural infusions as compared to preinfusion requirements (p less than 0.001). There were no instances of epidural scarring, respiratory depression, epidural infections, meningitis, or catheter blockage. One patient developed apparent drug tolerance and three patients required further catheter manipulations. This series strongly suggests that significant reductions in cancer pain can be obtained with few complications and a low morphine tolerance rate using chronic epidural morphine infusion. Anesthesiology and psychiatry input, along with temporary catheter infusion screening and quantitative pain evaluations using analog scales, are essential.