Epidural morphine injection followed by a steroid has been reported to be effective for the post-laminectomy pain ('failed back') syndrome. This double-blind, parallel study was undertaken to evaluate that mode of therapy. Twenty-two patients who had undergone at least one prior laminectomy, who were still symptomatic, were randomized to receive 50 mg of lidocaine epidurally with: (a) 75 mg triamcinolone diacetate (TR); or (b) 8 mg of preservative-free morphine (MP); or (c) both (TR and MP), at 1 month intervals for 3 consecutive months. The spinal interspace identified with the patients' pain complaint was the site of injection. For each treatment, patients were admitted to the Clinical Research Center for 24 h and their condition continuously monitored with a pulse oximeter and apnea monitor. Five to 7 patients in each group had pain relief for less than 1 month. No patient given morphine had pain relief for more than 1 month. Life-threatening ventilatory depression occurred in the group given triamcinolone and morphine. The use of morphine alone or combined with slow release triamcinolone does not appear to be appropriate for the treatment of the post-laminectomy pain syndrome.