BACKGROUND Plasma Beta endorphin (BE) is an endogenous peptide opioid derived form pro-opiomelanocortin. Although the role of plasma BE in pain regulation is unclear, plasma BE levels have been reported to correlate inversely with pain levels in cancer pain. OBJECTIVE To measure BE levels in patients with cancer before and after pain relief with different analgesic modalities, to evaluate the relationship between cancer pain, pain control and plasma BE levels. METHODS Prospective interventional study conducted in a university hospital setting. Patients with intractable pain due to upper abdominal visceral malignancies (pancreatic, liver and gastric cancers) agreed to participate in the study. Pain relief was provided with one of four primary modalities: 1) oral continuous release morphine, 2) thoracic epidural morphine boluses, 3) alcohol celiac plexus blocks or 4) interpleural infusions of bupivacaine. Blood samples for plasma BE levels were taken before treatment and at the time of maximal pain relief obtained with the particular treatment modality. Pain levels were determined using a 10 cm visual analogue scale. BE levels were measured by competitive radioimmunoassay using 125I Beta endorphin. RESULTS Average pain scores decreased from 7.3 +/- 1.27 (SD) before treatment to 1.2 +/- 1.18 (SD) after treatment (p<0.0001). Satisfactory pain relief was obtained with each of the four treatment modalities and was associated with a significant increase in plasma Beta levels. The mean plasma BE level for all groups before treatment was 18.9 +/- 5.4 pg/ ml (range 2.0 to 29.6 pg/ml) compared to 38.7 +/- 17.6 pg/ml (range 13.2 to 67.9 pg/ml) after pain relief (p < 0.0001). Plasma BE levels increased with improved pain control with each of the four analgesic modalities, including oral and epidural morphine. CONCLUSIONS Plasma BE levels increased with improved pain control in patients with upper abdominal gastrointestinal malignancies. Although the role of plasma BE in pain pathophysiology is unclear, it appears that pain relief per se, and not the analgesic technique, modulates plasma BE levels. This suggests that plasma BE levels may serve as an objective measure of cancer pain severity and corroborate the patient's report of pain relief.