It is uncertain whether there exists a nociceptive component in malignant nerve pain responsive to NSAIDs and opioids. 20 patients with malignant nerve pain were randomly assigned to treatment with naproxen 1500 mg versus slow-release morphine 60 mg daily during 1 week, followed by cross-over medication during the second week in a double-blind, double-dummy protocol. In the 16 evaluable patients, a significant (P < 0.05) reduction of 26% (S.E. +/- 7.9) in pain intensity was reached at day 7, compared to baseline pain. At day 7, significant pain relief of 32% (P < 0.05) was observed in the naproxen group, but not in the morphine group (21%, P = 0.14). Patients using morphine needed approximately twice as much paracetamol rescue than patients using naproxen. Additional pain relief could be observed in 4/9 patients with cross-over medication. These data support the concept of a nociceptive component in malignant nerve pain responding to NSAIDs and opioids, and favour the combination of both an anti-inflammatory drug and an opioid for symptomatic pain relief.