We report an open, uncontrolled study designed to assess the effects of subcutaneous (SC) morphine on dyspnea of terminal cancer. Twenty patients with dyspnea caused by restrictive respiratory failure received an SC dose of morphine of 5 mg (5 patients who were not receiving narcotics), or equivalent to 2.5 times their regular dose (15 patients who were receiving narcotics for pain). Dyspnea (D) and pain (15 cases) were measured before the dose and every 15 min for 150 min after the injection using a visual analog scale 0-100. Respiratory rate (RR), respiratory effort (RE) (score 1-6), arterial saturation of O2 (SO2) and end-tidal PACO2 were determined before and 45 min after SC morphine. D, RR, RE, SO2, and PACO2 were 68 +/- 32, 32 +/- 7; 3.5 +/- 1.8, 87 +/- 10, and 31 +/- 12, respectively, before SC morphine, and 34 +/- 25 (P less than 0.001), 31 +/- 9 (P:NS), 3.2 +/- 1.9 (P:NS), 86 +/- 11 (P:NS), and 33 +/- 9 (P:NS), respectively, 45 min after SC morphine. Nineteen of 20 patients (95%) reported improved dyspnea after morphine. We conclude that morphine appears to improve dyspnea without causing a significant deterioration in respiratory function in terminal cancer patients. Double-blind placebo controlled studies are needed in this population.